From c1e65c6a3113308b0597ba9702588d4762a3b621 Mon Sep 17 00:00:00 2001
From: Tomas Bystrican <tom.bystrican@gmail.com>
Date: Sat, 8 Aug 2026 18:31:20 +0930
Subject: [PATCH 1/5] Re-enable Seedance docs for faces-off API (API-103).

Document Seedance 2.0 and 2.5 with family-specific limits, replace public consent how-tos with faces-off guidance, and restore the guide in nav while keeping the face-consent page off-nav.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 api-reference/endpoint/video/queue.mdx        |   8 +-
 ar/api-reference/endpoint/video/queue.mdx     |   8 +-
 ar/guides/media/seedance-2-0.mdx              | 304 +++++++++-------
 de/api-reference/endpoint/video/queue.mdx     |   8 +-
 de/guides/media/seedance-2-0.mdx              | 324 ++++++++++--------
 docs.json                                     |   9 +
 ...6-08-08-api-103-seedance-faces-off-docs.md |  25 ++
 ...-api-103-seedance-faces-off-docs-design.md |  79 +++++
 es/api-reference/endpoint/video/queue.mdx     |   8 +-
 es/guides/media/seedance-2-0.mdx              | 308 ++++++++++-------
 fr/api-reference/endpoint/video/queue.mdx     |   8 +-
 fr/guides/media/seedance-2-0.mdx              | 300 +++++++++-------
 guides/media/seedance-2-0.mdx                 | 142 +++++---
 it/api-reference/endpoint/video/queue.mdx     |   8 +-
 it/guides/media/seedance-2-0.mdx              | 312 ++++++++++-------
 ko/api-reference/endpoint/video/queue.mdx     |   8 +-
 ko/guides/media/seedance-2-0.mdx              | 308 ++++++++++-------
 pt-BR/api-reference/endpoint/video/queue.mdx  |   8 +-
 pt-BR/guides/media/seedance-2-0.mdx           | 302 +++++++++-------
 zh/api-reference/endpoint/video/queue.mdx     |   8 +-
 zh/guides/media/seedance-2-0.mdx              | 302 +++++++++-------
 21 files changed, 1666 insertions(+), 1121 deletions(-)
 create mode 100644 docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md
 create mode 100644 docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md

diff --git a/api-reference/endpoint/video/queue.mdx b/api-reference/endpoint/video/queue.mdx
index fbee520d..c5c974c2 100644
--- a/api-reference/endpoint/video/queue.mdx
+++ b/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Call `/video/quote` to get a price estimate, then poll `/video/retrieve` with th
 
 Private models also return a `download_url` for the finished video. It is a short-lived delivery URL (a few retries are fine if a download drops); see the [Video Generation guide](/guides/media/video-generation#private-download-links) for details and optional `DELETE` for privacy.
 
-### Seedance 2.0
+### Seedance 2.0 & 2.5
 
-For `seedance-2-0-*` models (text-to-video, image-to-video, reference-to-video, plus the `-fast-*` variants), see the [Seedance 2.0 Guide](/guides/media/seedance-2-0) for the four-workflow model (Reference / Edit / Extend / Stitch), canonical prompt patterns, multimodal input limits, and pricing details.
+For `seedance-2-0-*` and `seedance-2-5-*` models (text-to-video, image-to-video, reference-to-video, plus the Seedance 2.0 `-fast-*` variants), see the [Seedance 2.0 & 2.5 Guide](/guides/media/seedance-2-0) for the four-workflow model (Reference / Edit / Extend / Stitch), family-specific multimodal limits, faces-off API behavior, and pricing details.
 
-### Seedance consent
+### Seedance faces-off API
 
-When a Seedance image- or reference-to-video request contains a human face, the API returns a `409 needs_consent` with the policy text, and you resubmit the same request with a `consents.seedance` attestation. See the [Seedance Face-Media Consent Guide](/guides/media/seedance-face-consent) for the full flow, the consent object, dedupe, and revocation.
+Public Seedance models do **not** support face-asset provisioning or face-consent attestation. Face-bearing inputs may be rejected upstream. For likeness / face-asset generation, use the Venice app or Studio. See the [Seedance guide](/guides/media/seedance-2-0) for details.
 
 ### Video upscaling
 
diff --git a/ar/api-reference/endpoint/video/queue.mdx b/ar/api-reference/endpoint/video/queue.mdx
index 0b982a72..bde1d794 100644
--- a/ar/api-reference/endpoint/video/queue.mdx
+++ b/ar/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ openapi: 'POST /video/queue'
 
 تُعيد النماذج الخاصة أيضًا `download_url` للفيديو النهائي. وهو عنوان URL للتسليم قصير الأمد (لا بأس بإعادة المحاولة بضع مرات في حال انقطع التنزيل)؛ راجع [دليل توليد الفيديو](/guides/media/video-generation#private-download-links) للتفاصيل و`DELETE` الاختيارية للخصوصية.
 
-### Seedance 2.0
+### Seedance 2.0 و 2.5
 
-بالنسبة لنماذج `seedance-2-0-*` (text-to-video، image-to-video، reference-to-video، بالإضافة إلى متغيرات `-fast-*`)، راجع [دليل Seedance 2.0](/guides/media/seedance-2-0) لنموذج التدفقات الأربعة (Reference / Edit / Extend / Stitch)، وأنماط الـ prompt القياسية، وحدود الإدخال متعدد الوسائط، وتفاصيل التسعير.
+لنماذج `seedance-2-0-*` و `seedance-2-5-*` (text-to-video و image-to-video و reference-to-video، بالإضافة إلى متغيّرات `-fast-*` لـ Seedance 2.0)، راجع [دليل Seedance 2.0 و 2.5](/guides/media/seedance-2-0) لنموذج سير العمل الأربعة (Reference / Edit / Extend / Stitch) وحدود الوسائط المتعددة حسب العائلة وسلوك واجهة faces-off وتفاصيل التسعير.
 
-### موافقة Seedance
+### واجهة Seedance بدون وجوه (faces-off)
 
-عندما يحتوي طلب image-to-video أو reference-to-video في Seedance على وجه بشري، تُعيد الـ API استجابة `409 needs_consent` مع نص السياسة، وتعيد إرسال الطلب نفسه مع شهادة `consents.seedance`. راجع [دليل موافقة Seedance لوسائط الوجوه](/guides/media/seedance-face-consent) للتدفق الكامل وكائن الموافقة وإزالة التكرار والإلغاء.
+نماذج Seedance العامة **لا** تدعم تجهيز أصول الوجه ولا إثبات موافقة الوجه. قد تُرفض المدخلات التي تحتوي على وجوه في المنبع. لتوليد likeness / أصول الوجه استخدم تطبيق Venice أو Studio. راجع [دليل Seedance](/guides/media/seedance-2-0).
 
 ### تكبير الفيديو
 
diff --git a/ar/guides/media/seedance-2-0.mdx b/ar/guides/media/seedance-2-0.mdx
index 9c7eacf0..a8281281 100644
--- a/ar/guides/media/seedance-2-0.mdx
+++ b/ar/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "أنشئ، حرّر، مدّد، ولصِّق مقاطع الفيديو مع Seedance 2.0 على Venice — تدفقات نص وصورة ومرجع إلى فيديو على endpoint واحد، مع أمثلة مشروحة."
-'og:title': "دليل Seedance 2.0 | وثائق Venice API"
-'og:description': "ولِّد وحرِّر ومدِّد وادمج مقاطع الفيديو باستخدام Seedance 2.0 عبر Venice API. أنماط سير العمل وحدود الإدخال متعدد الوسائط والتسعير وأمثلة عملية."
+title: "Seedance 2.0 و 2.5"
+description: "أنشئ وحرّر ومدّد والصِق مقاطع الفيديو مع Seedance 2.0 و 2.5 على Venice - تدفقات النص/الصورة/المرجع، وسلوك واجهة faces-off، وحدود الوسائط المتعددة حسب العائلة."
+'og:title': "دليل Seedance 2.0 و 2.5 | وثائق Venice API"
+'og:description': "ولِّد وحرِّر ومدِّد وادمج مقاطع الفيديو باستخدام Seedance 2.0 و 2.5 عبر Venice API. أنماط سير العمل وfaces-off والحدود والتسعير والأمثلة."
 ---
 
-Seedance 2.0 هو نموذج فيديو رائد متعدد الوسائط، متاح على Venice كعائلة من ثلاثة متغيّرات لتوليد الفيديو من النص والصورة والمراجع. متغيّر **reference-to-video** قويّ بشكل غير معتاد: نقطة نهاية واحدة ومعرّف نموذج واحد يتعاملان مع **أربعة سير عمل مميّزة** (Reference، Edit، Extend، Stitch) — ويُستنتج سير العمل من **شكل التعليمة**.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-يستعرض هذا الدليل المتغيّرات، وسير العمل الأربعة مع التعليمات المعيارية، وحدود الإدخال متعدد الوسائط، والتسعير، وأمثلة `curl` كاملة.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## المتغيّرات
+<Warning>
+**الوجوه غير مدعومة على واجهة Seedance العامة.** لا تشغّل Venice تجهيز أصول الوجه ولا إثبات موافقة الوجه (`consents.seedance` / `needs_consent`) لنماذج Seedance العامة. قد تُرفض المدخلات التي تحتوي على وجوه في المنبع كخطأ سياسة محتوى / مزوّد. لتوليد likeness / أصول الوجه استخدم تطبيق Venice أو Studio.
+</Warning>
+
+## Variants
 
-| معرّف النموذج | المتغيّر | دقّات الإخراج | ملاحظات |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | تعليمة نصية فقط |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | تثبيت بصورة الإطار الأول (واختياريًا الإطار الأخير) |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | حتى 9 صور مرجعية + 3 فيديوهات مرجعية + 3 مانحات صوتية. يدعم Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | فئة أسرع بجودة أقل |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | فئة أسرع بجودة أقل |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | فئة أسرع بجودة أقل؛ نفس مجموعة سير العمل |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-جميع المتغيّرات غير متزامنة. أرسل عبر `POST /api/v1/video/queue`، ثم استعلم بـ `POST /api/v1/video/retrieve` حتى يصبح جسم الاستجابة `video/mp4`. راجع [توليد الفيديو](/guides/media/video-generation) لتدفق الطابور العام.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## نموذج «نموذج واحد، أربعة سير عمل»
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-متغيّر reference-to-video (`seedance-2-0-reference-to-video` وشقيقه Fast) هو النموذج الأساسي ذاته يخدم أربع مهام مختلفة. **يستنتج النموذج المهمة من بادئة التعليمة وشكل مدخلاتك.** لا يوجد حقل `task` أو `workflow` — فبنية التعليمة هي ما يحدّد التوجيه.
+## The "one model, four workflows" model
 
-| سير العمل | ما يفعله | بادئة التعليمة | المدخلات |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | توليد فيديو جديد باستخدام الملفات المرجعية المرفوعة كمانحات للموضوع / الحركة / الأسلوب / الصوت | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | نص + ≥1 مرجع صورة أو فيديو (0–9 صور، 0–3 فيديوهات)، بالإضافة اختياريًا إلى ما يصل إلى 3 مانحات صوتية |
-| **Edit** | تعديل فيديو إدخال واحد مع الحفاظ على الباقي | `Strictly edit <Video 1>, changing its ...` | فيديو إدخال واحد + نص (صور اختيارية للتثبيت) |
-| **Extend** | تمديد أمامي / خلفي لمقطع واحد | `Extend <Video 1>, generate ...` | فيديو إدخال واحد + نص |
-| **Stitch** | دمج 2–3 مقاطع مع انتقالات مُولَّدة تلقائيًا | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2–3 فيديوهات إدخال + نص |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-**بنية التعليمة معيارية وحسّاسة لحالة الأحرف**: أقواس زاوية، حرف أول كبير، مسافة واحدة قبل الرقم — `<Video 1>`، `<Image 1>`، `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## أنماط سير العمل
+## Workflow patterns
 
-### سير العمل Reference
+### Reference workflow
 
-استخدم الملفات المرجعية المرفوعة كـ**مانحات** — للموضوع، والمشهد، والحركة، والأسلوب، ونبرة الصوت — لتوليد فيديو جديد كليًا.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**أنماط التعليمة المعيارية**:
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,23 +60,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**أمثلة**:
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (يجب اقتران المانحات الصوتية بمرجع صورة أو فيديو واحد على الأقل — الصوت وحده يُرفض)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### سير العمل Edit
+### Edit workflow
 
-عدّل فيديو إدخال واحد. **كل ما لا يُذكر صراحة في التعليمة يُحفظ.** استخدم هذا عندما تريد تغييرًا محليًا (استبدال موضوع، تغيير طقس/لون، إضافة/إزالة عنصر) بدلًا من فيديو جديد كليًا.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**نمط التعليمة المعياري**:
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**أنماط فرعية لتحكّم أدق**:
+**Sub-patterns for finer control**:
 
 ```
 Add Elements:
@@ -80,57 +89,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**أمثلة**:
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-يجمع المثال الأخير بين Edit ومرجع صورة — وهذا مسموح تمامًا، إذ يستخدم النموذج `<Image 1>` كمانح بصري للاستبدال.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### سير العمل Extend
+### Extend workflow
 
-استمر بمقطع واحد للأمام أو للخلف زمنيًا. **افتراضيًا يعيد Seedance المحتوى الجديد فقط** — وليس الإدخال الأصلي متصلًا بالامتداد. هذا مقصود لاستمرارية الانتقال؛ وإذا أردت الحفاظ على المقطع الأصلي إلى جانب الامتداد، فقل ذلك صراحة:
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← يحفظ الإدخال في البداية
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← يحفظ الإدخال في النهاية
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-التعامل مع الانتقال: يستخرج النموذج تلقائيًا إطارات الانتقال للمزج السلس، ولا تُعاد توليد الأجزاء الأصلية من فيديو الإدخال.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**أمثلة**:
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### سير العمل Stitch (إكمال المسار)
+### Stitch workflow (Track Completion)
 
-اربط 2–3 مقاطع إدخال بانتقالات مُولَّدة بالذكاء الاصطناعي. يجب ألّا تتجاوز المدة الإجمالية المجمّعة للإدخال 15 ثانية.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**نمط التعليمة المعياري**:
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**أمثلة**:
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-يقصّ النموذج تلقائيًا المقاطع الموصِلة عند نقاط الالتقاء لضمان الاستمرارية.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## صيغة التعليمة الموحّدة
+## Universal prompt formula
 
-عبر سير العمل الأربعة، الصيغة الموصى بها للتأليف هي:
+Across all four workflows, the recommended authoring formula is:
 
 ```
 Subject + Motion + Environment (Optional)
@@ -139,72 +148,70 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**: الأساس المنطقي — حدّد «مَن» يؤدّي «أي فعل»
-- **Environment + Aesthetics**: الخلفية المكانية، الإضاءة، الأسلوب البصري
-- **Camera**: نوع لقطة أو حركة صريحة
-- **Audio**: مؤثرات صوتية محيطة أو توجيه صوتي لإخراج غامر
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-وضع هذه فوق بادئة سير العمل (مثل `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) يُنتج أعلى جودة من المخرجات.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## حدود الإدخال متعدد الوسائط
+## Multimodal input limits
 
-القيم أدناه هي ما يقبله Venice API. الطلبات خارج هذه النطاقات تُرفض على طبقة المخطط برمز 400 قبل الوصول إلى الاستدلال.
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### الصور
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| القيد | القيمة |
-|---|---|
-| طرق الإدخال | URL (`http://`، `https://`) أو رابط بيانات Base64 (`data:image/...`) |
-| الصيغ | `.jpeg`، `.png`، `.webp`، `.bmp`، `.tiff`، `.gif`، `.heic`، `.heif` |
-| نسبة العرض إلى الارتفاع (W / H) | حصرية `(0.4, 2.5)` |
-| أدنى ضلع | ≥ 300 بكسل |
-| عدد الصور: I2V الإطار الأول | 1 |
-| عدد الصور: I2V الإطار الأول + الأخير | 2 |
-| **عدد الصور: R2V (V2 / Fast)** | **1 – 9** |
+### Shared media floors
 
-### الفيديوهات
-
-| القيد | القيمة |
-|---|---|
-| طرق الإدخال | URL (`http://`، `https://`) أو رابط بيانات Base64 (`data:video/...`) |
-| الصيغ | `.mp4`، `.mov` |
-| ترميزات الفيديو | H.264 / AVC، H.265 / HEVC |
-| ترميزات الصوت (في الحاوية) | AAC، MP3 |
-| **مدة كل مقطع** | `[2, 15]` ث (شامل) |
-| **الحد الأقصى لعدد المقاطع** | 3 (R2V / Stitch / Extend) |
-| **إجمالي المدة المجمّعة** | ≤ 15 ث عبر جميع المقاطع |
-| **حجم كل مقطع** | ≤ 50 ميغابايت |
-
-### الصوت
-
-| القيد | القيمة |
+| Constraint | Value |
 |---|---|
-| طرق الإدخال | URL (`http://`، `https://`) أو رابط بيانات Base64 (`data:audio/...`) |
-| الصيغ | `.wav`، `.mp3` |
-| **مدة كل مقطع** | `[2, 15]` ث |
-| **الحد الأقصى لعدد المقاطع** | 3 |
-| **إجمالي المدة المجمّعة** | ≤ 15 ث عبر جميع المقاطع |
-| **حجم كل مقطع** | ≤ 15 ميغابايت |
-
-الصوت المرجعي مدعوم على متغيّرات R2V فقط. تُمرَّر كل مدخلة إلى النموذج كعنصر محتوى بدور `role: "reference_audio"` تشير إليه التعليمة بـ `<Audio 1>`، `<Audio 2>`، `<Audio 3>` — ويستخدم النموذج كل مقطع لنبرة صوت أو مؤثرات صوتية أو موسيقى خلفية اعتمادًا على كيفية صياغة التعليمة. حقل `audio_url` المفرد القديم يُربط بنفس شكل المحتوى وهو الآن مكافئ لتمرير `reference_audio_urls` بعنصر واحد.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**لا يمكن أن يكون `reference_audio_urls` المدخل المرجعي الوحيد.** يتطلب النموذج مرجع صورة أو فيديو واحدًا على الأقل إلى جانب أي مانح صوتي. اقرن `reference_audio_urls` بـ `reference_image_urls` أو `reference_video_urls` أو `image_url` أو `video_url` — تُرفض الطلبات الصوتية فقط.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### حجم الطلب
+### Request size
 
-تقبل نقطة نهاية الطابور أجسام JSON حتى **35 ميغابايت**. روابط البيانات المضمّنة لفيديوهات كبيرة قد تتجاوز هذا — لـ Stitch متعدد المقاطع خصوصًا، فضّل الـ URLs على base64 المضمّن.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## التسعير
+## Pricing
 
-استدعِ `POST /api/v1/video/quote` للحصول على عرض سعر لشكل طلب معيّن قبل إرساله إلى `/video/queue`. نقطة نهاية الاقتباس هي المصدر المعتمد الوحيد؛ قد تتغيّر تفاصيل التسعير ولا ينبغي تخزينها مؤقتًا أو تكرارها على جانب العميل.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-عند تضمين فيديو(هات) مرجعية في الطلب، مرّر أيضًا `reference_video_total_duration` (مجموع كل مدد المقاطع المرجعية بالثواني) لتطابُق عرض السعر مع ما سيحاسب عليه `/video/queue`:
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,11 +228,11 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## أمثلة كاملة
+## Complete examples
 
-جميع الأمثلة تفترض ضبط `VENICE_API_KEY` في البيئة.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
-### من نص إلى فيديو
+### Text-to-video
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### من صورة إلى فيديو (الإطار الأول)
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-لا يقبل `seedance-2-0-image-to-video` (ومتغيّره Fast) حقل `aspect_ratio` — إذ تُشتق نسبة الإخراج تلقائيًا من أبعاد صورة الإدخال. تمرير الحقل يُعيد 400 مع رسالة *"This model does not support aspect_ratio"*. استخدم متغيّرات T2V أو R2V إذا كنت تحتاج تحكّمًا صريحًا في نسبة العرض إلى الارتفاع.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### سير عمل Reference — مانح موضوع
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### سير عمل Reference — مانح موضوع + صوت
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### سير عمل Edit
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### سير عمل Edit مع تثبيت بصورة
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend إلى الأمام
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -344,7 +385,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 مقاطع)
+### Stitch (3 clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### الاستعلام عن الاكتمال
+### Polling for completion
 
-بعد كل إرسال إلى الطابور، احفظ `queue_id` المُعاد واستعلم بـ `/video/retrieve` حتى يصبح جسم الاستجابة `video/mp4`:
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-تكون الاستجابة JSON (`{ "status": "queued" | "running" | "failed", ... }`) حتى تكتمل المهمة، عندها يتحوّل جسم الاستجابة إلى بايتات `video/mp4`. راجع [توليد الفيديو](/guides/media/video-generation) لنمط الاستعلام الكامل.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## استكشاف الأخطاء وإصلاحها
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-يجب أن تتضمن طلبات reference-to-video واحدًا على الأقل من `reference_image_urls` أو `reference_video_urls` أو `image_references` أو `video_references`. التوليد النصي البحت ليس سير عمل R2V صالحًا — استخدم `seedance-2-0-text-to-video` بدلًا من ذلك. لا يكفي `reference_audio_urls` وحده (راجع قسم الصوت أعلاه).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-يحدّد النموذج الفيديوهات المرجعية بـ 3. إن احتجت إلى مقاطع أكثر، نفّذ Stitch أولًا (3 → 1)، ثم استخدم الإخراج كمرجع لمتابعة لاحقة.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / المجموع `> 15s`
+### Prompt routes to the wrong workflow
 
-مدة كل مقطع `[2, 15]` ثانية **شاملاً**؛ ومجموع كل الفيديوهات المرجعية محدّد أيضًا بـ 15 ثانية. اقتص المقاطع على جانب العميل قبل الإرسال.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### تعليمة توجِّه إلى سير العمل الخاطئ
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-يُستنتج سير العمل من بنية التعليمة. أخطاء توجيه شائعة:
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- ترغب في **Extend** لكنك تكتب `Refer to ...` ← يعامل النموذج فيديوك كـ*مانح*، وليس كقماشة للاستمرار
-- ترغب في **Stitch** لكنك تكتب `Refer to ...` ← يختار النموذج أحدها كمانح، ويتجاهل الباقي
-- ترغب في **Edit** لكنك تكتب `Generate a video based on <Video 1>` ← غامض؛ قد يلجأ النموذج إلى Reference
+### Face / likeness inputs rejected
 
-استخدم البادئات المعيارية كما هي مكتوبة بالضبط: `Strictly edit <Video 1>, ...`، `Extend <Video 1>, ...`، `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### عرض السعر لا يطابق المبلغ المُجدوَل
+### Quote doesn't match the queued amount
 
-إذا ضمّنت فيديو مرجعيًا ولم تمرّر `reference_video_total_duration` إلى `/video/quote`، فقد يختلف عرض السعر عن المبلغ المُجدوَل. مرّر دومًا `reference_video_total_duration` (مجموع كل مدد المقاطع المرجعية بالثواني) عند وجود فيديوهات مرجعية.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## مراجع
+## References
 
-- نقطة نهاية طابور فيديو Venice: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- نقطة نهاية الاقتباس في Venice: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- دليل مرافق: [المرجع إلى الفيديو](/guides/media/reference-to-video) (يغطي Kling O3 + Grok Imagine R2V)
-- دليل مرافق: [توليد الفيديو](/guides/media/video-generation) (نظرة عامة على الطابور / الاستعلام)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/de/api-reference/endpoint/video/queue.mdx b/de/api-reference/endpoint/video/queue.mdx
index 00033679..1f19e992 100644
--- a/de/api-reference/endpoint/video/queue.mdx
+++ b/de/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Rufen Sie `/video/quote` auf, um einen Preisvoranschlag zu erhalten, und fragen
 
 Private Modelle geben außerdem eine `download_url` für das fertige Video zurück. Es handelt sich um eine kurzlebige Delivery-URL (einige Wiederholungen sind in Ordnung, wenn ein Download abbricht); Details und das optionale `DELETE` aus Datenschutzgründen finden Sie im [Video Generation Guide](/guides/media/video-generation#private-download-links).
 
-### Seedance 2.0
+### Seedance 2.0 & 2.5
 
-Für `seedance-2-0-*`-Modelle (text-to-video, image-to-video, reference-to-video plus die `-fast-*`-Varianten) siehe den [Seedance-2.0-Leitfaden](/guides/media/seedance-2-0) für das Vier-Workflow-Modell (Reference / Edit / Extend / Stitch), kanonische Prompt-Muster, multimodale Eingabegrenzen und Preisdetails.
+Für `seedance-2-0-*`- und `seedance-2-5-*`-Modelle (Text-to-Video, Image-to-Video, Reference-to-Video sowie die Seedance-2.0-`-fast-*`-Varianten) siehe den [Seedance-2.0-&-2.5-Guide](/guides/media/seedance-2-0) zu den vier Workflows (Reference / Edit / Extend / Stitch), familienspezifischen Multimodal-Limits, Faces-off-API-Verhalten und Preisen.
 
-### Seedance-Einwilligung
+### Seedance Faces-off API
 
-Wenn eine Seedance-Image- oder Reference-to-Video-Anfrage ein menschliches Gesicht enthält, gibt die API ein `409 needs_consent` mit dem Richtlinientext zurück, und Sie reichen dieselbe Anfrage erneut mit einer `consents.seedance`-Bestätigung ein. Den vollständigen Ablauf, das Consent-Objekt, Dedup und Widerruf finden Sie im [Seedance-Face-Media-Consent-Leitfaden](/guides/media/seedance-face-consent).
+Öffentliche Seedance-Modelle unterstützen **kein** Face-Asset-Provisioning und keine Face-Consent-Attestierung. Eingaben mit Gesichtern können upstream abgelehnt werden. Für Likeness-/Face-Asset-Generierung die Venice-App oder Studio nutzen. Details im [Seedance-Guide](/guides/media/seedance-2-0).
 
 ### Video-Upscaling
 
diff --git a/de/guides/media/seedance-2-0.mdx b/de/guides/media/seedance-2-0.mdx
index 175df530..a808e13b 100644
--- a/de/guides/media/seedance-2-0.mdx
+++ b/de/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "Generieren, bearbeiten, verlängern und stitchen Sie Videos mit Seedance 2.0 auf Venice — Text-, Image- und Reference-to-Video an einem Endpoint."
-'og:title': "Seedance 2.0-Guide | Venice API Docs"
-'og:description': "Videos mit Seedance 2.0 über die Venice-API generieren, editieren, verlängern und stitchen. Workflow-Muster, multimodale Limits, Preise und ausgearbeitete Beispiele."
+title: "Seedance 2.0 & 2.5"
+description: "Videos mit Seedance 2.0 und 2.5 auf Venice erzeugen, bearbeiten, verlängern und zusammenfügen - Text-/Bild-/Referenz-Workflows, Faces-off-API und familienspezifische Multimodal-Limits."
+'og:title': "Seedance 2.0 & 2.5 Guide | Venice API Docs"
+'og:description': "Videos mit Seedance 2.0 und 2.5 über die Venice API erzeugen. Workflows, Faces-off, Multimodal-Limits, Preise und Beispiele."
 ---
 
-Seedance 2.0 ist ein Flaggschiff-Multimodal-Videomodell, das in Venice als Familie aus drei Varianten für Text-, Image- und Referenz-getriebene Videogenerierung verfügbar ist. Die **Reference-to-Video**-Variante ist außergewöhnlich mächtig: ein einziger Endpoint und eine einzige Modell-ID handhaben **vier verschiedene Workflows** (Reference, Edit, Extend, Stitch) — der Workflow wird aus der **Form deines Prompts** abgeleitet.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-Dieser Guide führt durch die Varianten, die vier Workflows mit kanonischen Prompts, die Multimodal-Input-Limits, Preise und vollständige `curl`-Beispiele.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## Varianten
+<Warning>
+**Gesichter werden auf der öffentlichen Seedance-API nicht unterstützt.** Venice führt für öffentliche Seedance-Modelle kein Face-Asset-Provisioning und keine Face-Consent-Attestierung (`consents.seedance` / `needs_consent`) aus. Eingaben mit Gesichtern können upstream als Content-Policy-/Provider-Fehler abgelehnt werden. Für Likeness-/Face-Asset-Generierung die Venice-App oder Studio nutzen.
+</Warning>
+
+## Variants
 
-| Modell-ID | Variante | Output-Auflösungen | Hinweise |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Nur Text-Prompt |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Erst- (und optional Letzt-)Frame-Bildgrundlage |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Bis zu 9 Referenzbilder + 3 Referenzvideos + 3 Referenz-Audio-Donor. Treibt Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Schneller, niedrigerer Fidelity-Tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Schneller, niedrigerer Fidelity-Tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Schneller, niedrigerer Fidelity-Tier; selber Workflow-Satz |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-Alle Varianten sind async. Submit per `POST /api/v1/video/queue`, dann `POST /api/v1/video/retrieve` pollen, bis der Response-Body `video/mp4` ist. Siehe [Videogenerierung](/guides/media/video-generation) für den allgemeinen Queue-Flow.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## Das „Ein Modell, vier Workflows"-Modell
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-Die Reference-to-Video-Variante (`seedance-2-0-reference-to-video` und das Fast-Gegenstück) ist dasselbe zugrundeliegende Modell, das vier verschiedene Aufgaben bedient. **Das Modell leitet die Aufgabe aus dem Prompt-Präfix und der Form deiner Inputs ab.** Es gibt kein `task`- oder `workflow`-Feld – die Prompt-Syntax ist das Routing.
+## The "one model, four workflows" model
 
-| Workflow | Was es tut | Prompt-Präfix | Inputs |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Ein neues Video generieren, mit hochgeladenen Referenzdateien als Donor für Subjekt / Bewegung / Stil / Audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 Bild- ODER Video-Referenz (0–9 Bilder, 0–3 Videos), plus optional bis zu 3 Audio-Donor |
-| **Edit** | Ein einzelnes Eingabevideo modifizieren, den Rest erhalten | `Strictly edit <Video 1>, changing its ...` | 1 Eingabevideo + Text (Bilder optional als Grounding) |
-| **Extend** | Vorwärts-/Rückwärts-Verlängerung eines Clips | `Extend <Video 1>, generate ...` | 1 Eingabevideo + Text |
-| **Stitch** | 2–3 Clips mit automatisch generierten Übergängen verbinden | `<Video 1> + <Transition-Beschreibung> + followed by <Video 2> + ...` | 2–3 Eingabevideos + Text |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-Die **Prompt-Syntax ist kanonisch und case-sensitiv**: spitze Klammern, Großbuchstabe am Anfang, ein Leerzeichen vor der Zahl – `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow-Muster
+## Workflow patterns
 
-### Reference-Workflow
+### Reference workflow
 
-Nutze die hochgeladenen Referenzdateien als **Donor** – Subjekt, Szene, Bewegung, Stil, Stimmtimbre –, um ein völlig neues Video zu generieren.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**Kanonische Prompt-Muster**:
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,160 +60,158 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Beispiele**:
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (Audio-Donor müssen mit mindestens einer Bild- oder Video-Referenz gepaart sein – nur Audio wird abgelehnt)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Edit-Workflow
+### Edit workflow
 
-Modifiziere ein einzelnes Eingabevideo. **Alles, was nicht explizit im Prompt genannt ist, bleibt erhalten.** Nutze dies, wenn du eine lokale Änderung möchtest (Subjekttausch, Wetter-/Farbänderung, Element-Add/Remove) statt eines komplett neuen Videos.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**Kanonisches Prompt-Muster**:
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-Muster für feinere Kontrolle**:
+**Sub-patterns for finer control**:
 
 ```
-Elemente hinzufügen:
+Add Elements:
   At [timestamp / timing] and [spatial location] of <Video 1>, add [description of intended element].
 
-Elemente entfernen:
+Remove Elements:
   Remove [element to be deleted] from <Video 1>, keeping the rest of the video content unchanged.
 
-Elemente ändern:
+Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Beispiele**:
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-Das letzte Beispiel kombiniert Edit mit einer Bildreferenz – völlig legitim, das Modell nutzt `<Image 1>` als visuellen Donor für den Austausch.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Extend-Workflow
+### Extend workflow
 
-Setze einen einzelnen Clip zeitlich vorwärts oder rückwärts fort. **Standardmäßig liefert Seedance nur den neuen Inhalt** – nicht das Original-Input verkettet mit der Verlängerung. Das ist für die Übergangskontinuität gewollt; soll der Eingabeclip neben der Verlängerung erhalten bleiben, sage es explizit:
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
-Extend <Video 1>, generate [Beschreibung des Verlängerungsinhalts]
-Extend <Video 1> backward, [Beschreibung des Verlängerungsinhalts]
-Extend <Video 1>, start with <Video 1>, then [Beschreibung des Verlängerungsinhalts]      ← Input am Anfang erhalten
-Extend <Video 1> backward, [Beschreibung], and then end with <Video 1>                    ← Input am Ende erhalten
+Extend <Video 1>, generate [description of extended content]
+Extend <Video 1> backward, [description of extended content]
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-Übergangsbehandlung: Das Modell extrahiert automatisch die Übergangsframes für nahtloses Blending; die Originalsegmente des Eingabevideos werden nicht neu generiert.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**Beispiele**:
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch-Workflow (Track-Completion)
+### Stitch workflow (Track Completion)
 
-Verbinde 2–3 Eingangsclips mit KI-generierten Übergängen. Die kombinierte Eingangsdauer muss ≤ 15 s sein.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**Kanonisches Prompt-Muster**:
+**Canonical prompt pattern**:
 
 ```
-<Video 1> + [Transition-Beschreibung] + followed by <Video 2> [+ [Transition-Beschreibung] + followed by <Video 3>]
+<Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Beispiele**:
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-Das Modell trimmt verbindende Segmente an den Übergängen automatisch für Kontinuität.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## Universelle Prompt-Formel
+## Universal prompt formula
 
-Über alle vier Workflows hinweg lautet die empfohlene Schreibformel:
+Across all four workflows, the recommended authoring formula is:
 
 ```
-Subjekt + Bewegung + Umgebung (optional)
-       + Kamerabewegung / Schnitt (optional)
-       + Ästhetik-Beschreibung (optional)
-       + Audio (optional)
+Subject + Motion + Environment (Optional)
+       + Camera Movement / Cut (Optional)
+       + Aesthetic Description (Optional)
+       + Audio (Optional)
 ```
 
-- **Subjekt + Bewegung**: das logische Fundament – definiere „Wer" führt „Welche Aktion" aus
-- **Umgebung + Ästhetik**: räumlicher Hintergrund, Beleuchtung, Bildstil
-- **Kamera**: expliziter Shot-Typ oder Bewegung
-- **Audio**: Ambient-Soundeffekte oder Stimmrichtung für immersiven Output
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-Wenn du das auf ein Workflow-Präfix legst (z. B. `Strictly edit <Video 1>, changing its <Subjekt + Bewegung + Umgebung + ...>`), bekommst du die qualitativ besten Ergebnisse.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## Multimodale Input-Limits
+## Multimodal input limits
 
-Die unten genannten Werte sind das, was die Venice-API akzeptiert. Anfragen außerhalb dieser Bereiche werden auf der Schema-Ebene mit 400 abgelehnt, bevor sie die Inferenz erreichen.
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### Bilder
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| Constraint | Wert |
-|---|---|
-| Input-Methoden | URL (`http://`, `https://`) oder Base64-Data-URL (`data:image/...`) |
-| Formate | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Aspect Ratio (B / H) | exklusiv `(0.4, 2.5)` |
-| Mindestseite | ≥ 300 px |
-| Bildanzahl: I2V Erst-Frame | 1 |
-| Bildanzahl: I2V Erst- + Letzt-Frame | 2 |
-| **Bildanzahl: R2V (V2 / Fast)** | **1 – 9** |
+### Shared media floors
 
-### Videos
-
-| Constraint | Wert |
-|---|---|
-| Input-Methoden | URL (`http://`, `https://`) oder Base64-Data-URL (`data:video/...`) |
-| Formate | `.mp4`, `.mov` |
-| Video-Codecs | H.264 / AVC, H.265 / HEVC |
-| Audio-Codecs (im Container) | AAC, MP3 |
-| **Dauer pro Clip** | `[2, 15]` s (inklusiv) |
-| **Max. Clip-Anzahl** | 3 (R2V / Stitch / Extend) |
-| **Gesamtdauer kombiniert** | ≤ 15 s über alle Clips |
-| **Größe pro Clip** | ≤ 50 MB |
-
-### Audio
-
-| Constraint | Wert |
+| Constraint | Value |
 |---|---|
-| Input-Methoden | URL (`http://`, `https://`) oder Base64-Data-URL (`data:audio/...`) |
-| Formate | `.wav`, `.mp3` |
-| **Dauer pro Clip** | `[2, 15]` s |
-| **Max. Clip-Anzahl** | 3 |
-| **Gesamtdauer kombiniert** | ≤ 15 s über alle Clips |
-| **Größe pro Clip** | ≤ 15 MB |
-
-Referenz-Audio wird nur in den R2V-Varianten unterstützt. Jeder Eintrag wird an das Modell als Content-Item mit `role: "reference_audio"` weitergegeben, das der Prompt als `<Audio 1>`, `<Audio 2>`, `<Audio 3>` adressiert – das Modell nutzt jeden Clip je nach Prompt-Rahmen für Stimmtimbre, Soundeffekte oder Hintergrundmusik. Das veraltete Singular-Feld `audio_url` mappt auf dieselbe Content-Form und ist nun äquivalent zur Übergabe eines einzelelementigen `reference_audio_urls`.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` kann nicht der einzige Referenz-Input sein.** Das Modell verlangt mindestens eine Bild- oder Videoreferenz neben jedem Audio-Donor. Paare `reference_audio_urls` mit `reference_image_urls`, `reference_video_urls`, `image_url` oder `video_url` – reine Audio-Submissions werden abgelehnt.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### Request-Größe
+### Request size
 
-Der Queue-Endpoint akzeptiert JSON-Bodies bis zu **35 MB**. Inline-Data-URLs für große Videos können das überschreiten – besonders bei Multi-Clip-Stitch URLs gegenüber Inline-Base64 bevorzugen.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## Preise
+## Pricing
 
-Rufe `POST /api/v1/video/quote` auf, um ein Quote für eine bestimmte Request-Form zu erhalten, bevor du es an `/video/queue` schickst. Der Quote-Endpoint ist die einzige maßgebliche Quelle; Pricing-Details können sich ändern und sollten nicht client-seitig gecached oder dupliziert werden.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-Wenn Referenzvideos Teil des Requests sind, übergib auch `reference_video_total_duration` (die Summe aller Referenzclip-Dauern in Sekunden), damit das Quote dem entspricht, was `/video/queue` abrechnet:
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,11 +228,11 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Vollständige Beispiele
+## Complete examples
 
-Alle Beispiele setzen voraus, dass `VENICE_API_KEY` in der Umgebung gesetzt ist.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
-### Text-to-Video
+### Text-to-video
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Image-to-Video (Erst-Frame)
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video` (und seine Fast-Variante) **akzeptieren `aspect_ratio` nicht** – das Output-Aspect-Ratio wird automatisch aus den Dimensionen des Eingabebilds abgeleitet. Wird das Feld übergeben, kommt 400 mit *„This model does not support aspect_ratio"*. Verwende die T2V- oder R2V-Varianten, wenn du explizite Aspect-Ratio-Kontrolle brauchst.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Reference-Workflow — Subjekt-Donor
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Reference-Workflow — Subjekt + Audio-Donor
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit-Workflow
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit-Workflow mit Bild-Grounding
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend (vorwärts)
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -344,7 +385,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 Clips)
+### Stitch (3 clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling auf Fertigstellung
+### Polling for completion
 
-Speichere nach jedem Queue-Submit die zurückgegebene `queue_id` und pollte `/video/retrieve`, bis der Response-Body `video/mp4` ist:
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-Die Antwort ist JSON (`{ "status": "queued" | "running" | "failed", ... }`), bis der Job abgeschlossen ist; dann wechselt der Response-Body auf `video/mp4`-Bytes. Siehe [Videogenerierung](/guides/media/video-generation) für das vollständige Polling-Muster.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## Fehlerbehebung
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-Reference-to-Video-Submissions müssen mindestens eines von `reference_image_urls`, `reference_video_urls`, `image_references` oder `video_references` enthalten. Reine Text-Generierung ist kein gültiger R2V-Workflow – verwende stattdessen `seedance-2-0-text-to-video`. `reference_audio_urls` allein reicht nicht (siehe Audio-Abschnitt oben).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-Das Modell deckelt Referenzvideos bei 3. Wenn du mehr Clips brauchst, mach zuerst einen Stitch (3 → 1) und nutze den Output als Referenz für einen Folge-Job.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / Aggregat `> 15s`
+### Prompt routes to the wrong workflow
 
-Die Dauer pro Clip ist `[2, 15]` Sekunden **inklusiv**; die Summe über alle Referenzvideos ist ebenfalls auf 15 Sekunden gedeckelt. Clips vor dem Submit client-seitig trimmen.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### Prompt routet auf den falschen Workflow
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-Der Workflow wird aus der Prompt-Syntax abgeleitet. Häufige Fehlrouten:
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- **Extend** gemeint, aber `Refer to ...` geschrieben → Modell behandelt dein Video als *Donor*, nicht als fortzusetzende Canvas
-- **Stitch** gemeint, aber `Refer to ...` geschrieben → Modell wählt einen als Donor, ignoriert die anderen
-- **Edit** gemeint, aber `Generate a video based on <Video 1>` geschrieben → mehrdeutig; Modell fällt evtl. auf Reference zurück
+### Face / likeness inputs rejected
 
-Nutze die kanonischen Präfixe exakt wie geschrieben: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### Quote stimmt nicht mit dem Queue-Betrag überein
+### Quote doesn't match the queued amount
 
-Wenn du ein Referenzvideo eingebunden, aber `reference_video_total_duration` nicht an `/video/quote` übergeben hast, können Quote und Queue-Betrag abweichen. Bei vorhandenen Referenzvideos immer `reference_video_total_duration` (Summe aller Referenzclip-Dauern in Sekunden) übergeben.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## Referenzen
+## References
 
-- Venice Video-Queue-Endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice Quote-Endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Begleitender Guide: [Reference to Video](/guides/media/reference-to-video) (deckt Kling O3 + Grok Imagine R2V ab)
-- Begleitender Guide: [Videogenerierung](/guides/media/video-generation) (Queue-/Polling-Übersicht)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/docs.json b/docs.json
index 834eed28..30f53574 100644
--- a/docs.json
+++ b/docs.json
@@ -70,6 +70,7 @@
                   "guides/media/image-editing",
                   "guides/media/image-upscaling",
                   "guides/media/video-generation",
+                  "guides/media/seedance-2-0",
                   "guides/media/reference-to-video",
                   "guides/media/music-and-sound-effects",
                   "guides/media/text-to-speech",
@@ -384,6 +385,7 @@
                   "pt-BR/guides/media/image-editing",
                   "pt-BR/guides/media/image-upscaling",
                   "pt-BR/guides/media/video-generation",
+                  "pt-BR/guides/media/seedance-2-0",
                   "pt-BR/guides/media/reference-to-video",
                   "pt-BR/guides/media/music-and-sound-effects",
                   "pt-BR/guides/media/text-to-speech",
@@ -698,6 +700,7 @@
                   "ar/guides/media/image-editing",
                   "ar/guides/media/image-upscaling",
                   "ar/guides/media/video-generation",
+                  "ar/guides/media/seedance-2-0",
                   "ar/guides/media/reference-to-video",
                   "ar/guides/media/music-and-sound-effects",
                   "ar/guides/media/text-to-speech",
@@ -1012,6 +1015,7 @@
                   "it/guides/media/image-editing",
                   "it/guides/media/image-upscaling",
                   "it/guides/media/video-generation",
+                  "it/guides/media/seedance-2-0",
                   "it/guides/media/reference-to-video",
                   "it/guides/media/music-and-sound-effects",
                   "it/guides/media/text-to-speech",
@@ -1326,6 +1330,7 @@
                   "de/guides/media/image-editing",
                   "de/guides/media/image-upscaling",
                   "de/guides/media/video-generation",
+                  "de/guides/media/seedance-2-0",
                   "de/guides/media/reference-to-video",
                   "de/guides/media/music-and-sound-effects",
                   "de/guides/media/text-to-speech",
@@ -1640,6 +1645,7 @@
                   "es/guides/media/image-editing",
                   "es/guides/media/image-upscaling",
                   "es/guides/media/video-generation",
+                  "es/guides/media/seedance-2-0",
                   "es/guides/media/reference-to-video",
                   "es/guides/media/music-and-sound-effects",
                   "es/guides/media/text-to-speech",
@@ -1954,6 +1960,7 @@
                   "fr/guides/media/image-editing",
                   "fr/guides/media/image-upscaling",
                   "fr/guides/media/video-generation",
+                  "fr/guides/media/seedance-2-0",
                   "fr/guides/media/reference-to-video",
                   "fr/guides/media/music-and-sound-effects",
                   "fr/guides/media/text-to-speech",
@@ -2268,6 +2275,7 @@
                   "zh/guides/media/image-editing",
                   "zh/guides/media/image-upscaling",
                   "zh/guides/media/video-generation",
+                  "zh/guides/media/seedance-2-0",
                   "zh/guides/media/reference-to-video",
                   "zh/guides/media/music-and-sound-effects",
                   "zh/guides/media/text-to-speech",
@@ -2582,6 +2590,7 @@
                   "ko/guides/media/image-editing",
                   "ko/guides/media/image-upscaling",
                   "ko/guides/media/video-generation",
+                  "ko/guides/media/seedance-2-0",
                   "ko/guides/media/reference-to-video",
                   "ko/guides/media/music-and-sound-effects",
                   "ko/guides/media/text-to-speech",
diff --git a/docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md b/docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md
new file mode 100644
index 00000000..1c7d99c6
--- /dev/null
+++ b/docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md
@@ -0,0 +1,25 @@
+# API-103 Seedance faces-off docs Implementation Plan
+
+> **For agentic workers:** Implement task-by-task. Steps use checkbox syntax.
+
+**Goal:** Re-enable Seedance public docs for faces-off API (2.0 + 2.5 limits), leave face-consent off-nav unchanged, fix queue reference blurbs.
+
+**Architecture:** Mintlify MDX guides + `docs.json` nav; content sourced from outerface Seedance model defs / ENG-950 product rules.
+
+**Tech Stack:** Mintlify, MDX, JSON nav
+
+## Global Constraints
+
+- Keep URL `guides/media/seedance-2-0`; retitle for 2.0 + 2.5
+- Face-consent pages: no content edits; stay off-nav
+- Do not edit `swagger.yaml`
+- Split 2.0 vs 2.5 multimodal limits; never present 2.0 caps as applying to 2.5
+- Full i18n for guide + queue blurb + nav
+
+## Tasks
+
+- [ ] Task 1: Update English `guides/media/seedance-2-0.mdx`
+- [ ] Task 2: Update English `api-reference/endpoint/video/queue.mdx`
+- [ ] Task 3: Re-enable Seedance guide in all `docs.json` locale navs
+- [ ] Task 4: Port guide + queue changes to ar/de/es/fr/it/ko/pt-BR/zh
+- [ ] Task 5: Grep for remaining live consent how-tos; sanity-check nav
diff --git a/docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md b/docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md
new file mode 100644
index 00000000..d226f2cc
--- /dev/null
+++ b/docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md
@@ -0,0 +1,79 @@
+# API-103 — Seedance faces-off public docs (re-enable)
+
+**Date:** 2026-08-08  
+**Status:** Approved  
+**Repo:** `veniceai/api-docs`  
+**Linear:** [API-103](https://linear.app/venice/issue/API-103)  
+**Depends on:** [ENG-950](https://linear.app/venice/issue/ENG-950) (hard-cut + listed no-face API twins)
+
+## Goal
+
+Public Venice API docs match faces-off Seedance API behavior after ENG-950, document Seedance **2.0 and 2.5** (with correct multimodal limits), and re-enable the Seedance guide in the public nav.
+
+## Decisions
+
+| Topic | Decision |
+| --- | --- |
+| Face-consent page | Keep file + locale copies; **leave content unchanged**; **stay off-nav** |
+| Seedance guide | Retitle for 2.0 + 2.5; keep URL `guides/media/seedance-2-0` |
+| Nav re-enable | On this docs PR merge (do not wait for a second PR) |
+| OpenAPI / `swagger.yaml` | Out of scope |
+| Queue API reference | Replace “Seedance consent” blurb with faces-off note + link to Seedance guide (all locales) |
+| Locales | Full i18n: EN + all locale copies of guide + queue blurb + `docs.json` nav |
+| Enhanced / Mini / 1.5 Pro | Out of this pass |
+
+## Faces-off messaging (required)
+
+Public docs must state:
+
+- Public API Seedance does **not** run face-asset / consent attestation (`consents.seedance` / `needs_consent` reseller path).
+- Face-bearing media may fail upstream (content-policy / provider rejection).
+- Likeness / face-asset generation: use the Venice app / Studio.
+- Do **not** link `seedance-face-consent` as a how-to for API callers.
+
+## Guide content changes
+
+File: `guides/media/seedance-2-0.mdx` (+ `ar|de|es|fr|it|ko|pt-BR|zh` copies)
+
+1. Title / description → Seedance 2.0 **and** 2.5.
+2. Variants table: keep 2.0 + Fast; add `seedance-2-5-{text,image,reference}-to-video`.
+3. Note R2V four-workflow model applies to 2.0 and 2.5 R2V.
+4. Prominent faces-off callout near the top (after intro / variants).
+5. **Split multimodal limits** by family (do not reuse a single 2.0 table for 2.5).
+
+### Limits (from outerface model defs)
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+| --- | --- | --- |
+| R2V reference images | 1–9 | 1–30 |
+| R2V reference videos | ≤3; combined ≤15s; per-clip 2–15s | ≤10; combined ≤30s; per-clip 2–30s |
+| R2V reference audio | ≤3; combined ≤15s | ≤10; combined ≤30s |
+| Output duration | 4–15s | 4–30s (default 10) |
+| Resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Shared floors | short side ≥300px; aspect (0.4, 2.5); mime/size as today | same floors; max image bytes 30MB where defined |
+
+6. Keep most examples on 2.0 IDs; add ≥1 Seedance 2.5 example (longer duration and/or higher ref counts).
+
+## Nav
+
+`docs.json` — under Image, Video & Audio (every language), insert `…/guides/media/seedance-2-0` after `video-generation`.  
+Do **not** add `seedance-face-consent`.
+
+## API reference
+
+`api-reference/endpoint/video/queue.mdx` (+ locales): replace “### Seedance consent” with faces-off note linking to the Seedance guide. Keep “### Seedance 2.0” section; extend wording to 2.0 / 2.5 where it only names 2.0.
+
+## Out of scope
+
+- Rewriting or deleting `seedance-face-consent*`
+- `swagger.yaml` / OpenAPI sync
+- Video harness Seedance face-rule cleanup
+- Mini / Enhanced / 1.5 Pro public docs
+
+## Acceptance
+
+- No **nav-linked** public page instructs API callers to submit Seedance face consent.
+- Seedance guide documents faces-off + 2.0 vs 2.5 limits accurately.
+- Seedance guide is in public nav for all locales.
+- Face-consent pages remain off-nav with unchanged body.
+- Queue reference no longer points readers at the consent how-to as current API behavior.
diff --git a/es/api-reference/endpoint/video/queue.mdx b/es/api-reference/endpoint/video/queue.mdx
index 32b0c08d..e4448fea 100644
--- a/es/api-reference/endpoint/video/queue.mdx
+++ b/es/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Llama a `/video/quote` para obtener una estimación de precio y, a continuación
 
 Los modelos privados también devuelven un `download_url` para el vídeo terminado. Es una URL de entrega de corta duración (algunos reintentos están bien si una descarga se interrumpe); consulta la [guía de generación de vídeo](/guides/media/video-generation#private-download-links) para conocer los detalles y el `DELETE` opcional para privacidad.
 
-### Seedance 2.0
+### Seedance 2.0 y 2.5
 
-Para los modelos `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*`), consulta la [Guía de Seedance 2.0](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), patrones canónicos de prompt, límites de entrada multimodal y detalles de precios.
+Para los modelos `seedance-2-0-*` y `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*` de Seedance 2.0), consulta la [Guía de Seedance 2.0 y 2.5](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), límites multimodales por familia, comportamiento faces-off de la API y detalles de precios.
 
-### Consentimiento de Seedance
+### Seedance API sin rostros (faces-off)
 
-Cuando una solicitud de image- o reference-to-video de Seedance contiene un rostro humano, la API devuelve un `409 needs_consent` con el texto de la política, y debes reenviar la misma solicitud con una atestación `consents.seedance`. Consulta la [Guía de consentimiento para contenido de rostro de Seedance](/guides/media/seedance-face-consent) para conocer el flujo completo, el objeto de consentimiento, deduplicación y revocación.
+Los modelos Seedance públicos **no** admiten aprovisionamiento de face assets ni atestación de consentimiento facial. Las entradas con rostros pueden ser rechazadas upstream. Para generación con likeness / face assets, usa la app o Studio de Venice. Consulta la [guía de Seedance](/guides/media/seedance-2-0).
 
 ### Escalado de vídeo
 
diff --git a/es/guides/media/seedance-2-0.mdx b/es/guides/media/seedance-2-0.mdx
index 3d3b58ae..7afa70ef 100644
--- a/es/guides/media/seedance-2-0.mdx
+++ b/es/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "Genera, edita, extiende y une vídeos con Seedance 2.0 en Venice — flujos de text, image y reference-to-video en un mismo endpoint, con ejemplos resueltos."
-'og:title': "Seedance 2.0 Guide | Venice API Docs"
-'og:description': "Genera, edita, extiende y une vídeos con Seedance 2.0 a través de la API de Venice. Patrones de flujo, límites multimodales, precios y ejemplos resueltos."
+title: "Seedance 2.0 y 2.5"
+description: "Genera, edita, amplía y une vídeos con Seedance 2.0 y 2.5 en Venice - flujos texto/imagen/referencia, comportamiento faces-off de la API y límites multimodales por familia."
+'og:title': "Guía Seedance 2.0 y 2.5 | Venice API Docs"
+'og:description': "Genera, edita, amplía y une vídeos con Seedance 2.0 y 2.5 vía la API de Venice. Patrones de flujo, faces-off, límites multimodales, precios y ejemplos."
 ---
 
-Seedance 2.0 es un modelo multimodal de vídeo insignia expuesto en Venice como una familia de tres variantes para generación de vídeo a partir de texto, imagen y referencia. La variante **reference-to-video** es excepcionalmente potente: un único endpoint y un único ID de modelo gestionan **cuatro flujos distintos** (Reference, Edit, Extend, Stitch) — el flujo se infiere de la **forma de tu prompt**.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-Esta guía recorre las variantes, los cuatro flujos con sus prompts canónicos, los límites de entrada multimodal, los precios y ejemplos completos con `curl`.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## Variantes
+<Warning>
+**Los rostros no están soportados en la API pública de Seedance.** Venice no ejecuta aprovisionamiento de face assets ni atestación de consentimiento facial (`consents.seedance` / `needs_consent`) para los modelos Seedance públicos. Las entradas con rostros pueden rechazarse upstream como error de política de contenido / proveedor. Para generación con likeness / face assets, usa la app o Studio de Venice.
+</Warning>
+
+## Variants
 
-| ID del modelo | Variante | Resoluciones de salida | Notas |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Solo prompt de texto |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Anclaje de primer frame (y opcionalmente último) con imagen |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Hasta 9 imágenes de referencia + 3 vídeos de referencia + 3 donantes de audio. Impulsa Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Nivel más rápido y de menor fidelidad |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Nivel más rápido y de menor fidelidad |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Nivel más rápido y de menor fidelidad; mismo conjunto de flujos |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-Todas las variantes son asíncronas. Envía mediante `POST /api/v1/video/queue` y luego haz polling de `POST /api/v1/video/retrieve` hasta que el cuerpo de la respuesta sea `video/mp4`. Consulta [Generación de vídeo](/guides/media/video-generation) para el flujo general de cola.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## El modelo "un modelo, cuatro flujos"
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-La variante reference-to-video (`seedance-2-0-reference-to-video` y su hermana Fast) es el mismo modelo subyacente sirviendo cuatro tareas diferentes. **El modelo infiere la tarea a partir del prefijo del prompt y la forma de tus entradas.** No hay un campo `task` o `workflow` — la sintaxis del prompt es el enrutamiento.
+## The "one model, four workflows" model
 
-| Flujo | Qué hace | Prefijo del prompt | Entradas |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Genera un vídeo nuevo usando los archivos de referencia subidos como donantes de sujeto / movimiento / estilo / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Texto + ≥1 referencia de imagen O vídeo (0-9 imágenes, 0-3 vídeos), más opcionalmente hasta 3 donantes de audio |
-| **Edit** | Modifica un único vídeo de entrada preservando el resto | `Strictly edit <Video 1>, changing its ...` | 1 vídeo de entrada + texto (imágenes como anclaje opcional) |
-| **Extend** | Extensión hacia adelante / atrás de un clip | `Extend <Video 1>, generate ...` | 1 vídeo de entrada + texto |
-| **Stitch** | Une 2-3 clips con transiciones auto-generadas | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 vídeos de entrada + texto |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-La **sintaxis del prompt es canónica y sensible a mayúsculas**: corchetes angulares, primera letra en mayúscula, un único espacio antes del número — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Patrones de flujo
+## Workflow patterns
 
-### Flujo Reference
+### Reference workflow
 
-Usa los archivos de referencia subidos como **donantes** — sujeto, escena, movimiento, estilo, timbre vocal — para generar un vídeo completamente nuevo.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**Patrones canónicos de prompt**:
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,23 +60,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Ejemplos**:
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (los donantes de audio deben emparejarse con al menos una referencia de imagen o vídeo — solo audio se rechaza)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Flujo Edit
+### Edit workflow
 
-Modifica un único vídeo de entrada. **Cualquier cosa no nombrada explícitamente en el prompt se preserva.** Úsalo cuando quieras un cambio localizado (cambio de sujeto, cambio de clima/color, añadir/eliminar elemento) en lugar de un vídeo totalmente nuevo.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**Patrón canónico de prompt**:
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patrones para control más fino**:
+**Sub-patterns for finer control**:
 
 ```
 Add Elements:
@@ -80,131 +89,129 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Ejemplos**:
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-El último ejemplo combina Edit con una referencia de imagen — totalmente legal, el modelo usa `<Image 1>` como donante visual para el reemplazo.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Flujo Extend
+### Extend workflow
 
-Continúa un único clip hacia adelante o hacia atrás en el tiempo. **Por defecto, Seedance devuelve solo el contenido nuevo** — no el clip de entrada original concatenado con la extensión. Esto es por diseño, para continuidad de transición; si quieres preservar el clip de entrada junto con la extensión, dilo explícitamente:
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserva el input al principio
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserva el input al final
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-Gestión de transiciones: el modelo extrae automáticamente los frames de transición para una mezcla sin costuras, y los segmentos originales del vídeo de entrada no se re-generan.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**Ejemplos**:
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Flujo Stitch (Track Completion)
+### Stitch workflow (Track Completion)
 
-Conecta 2-3 clips de entrada con transiciones generadas por IA. La duración total combinada de entrada debe ser ≤ 15 s.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**Patrón canónico de prompt**:
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Ejemplos**:
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-El modelo recorta automáticamente los segmentos de conexión en los puntos de unión para la continuidad.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## Fórmula universal de prompt
+## Universal prompt formula
 
-En los cuatro flujos, la fórmula de autoría recomendada es:
+Across all four workflows, the recommended authoring formula is:
 
 ```
-Sujeto + Movimiento + Entorno (opcional)
-       + Movimiento de cámara / Corte (opcional)
-       + Descripción estética (opcional)
-       + Audio (opcional)
+Subject + Motion + Environment (Optional)
+       + Camera Movement / Cut (Optional)
+       + Aesthetic Description (Optional)
+       + Audio (Optional)
 ```
 
-- **Sujeto + Movimiento**: la base lógica — define "Quién" está realizando "Qué acción"
-- **Entorno + Estética**: fondo espacial, iluminación, estilo visual
-- **Cámara**: tipo o movimiento explícito de toma
-- **Audio**: efectos de sonido ambientales o dirección vocal para salida inmersiva
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-Poner esta capa encima de un prefijo de flujo (p. ej., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produce las salidas de mayor calidad.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## Límites de entrada multimodal
+## Multimodal input limits
 
-Los valores siguientes son lo que acepta la API de Venice. Las solicitudes fuera de estos rangos se rechazan en la capa de schema con un 400 antes de llegar a la inferencia.
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### Imágenes
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| Restricción | Valor |
-|---|---|
-| Métodos de entrada | URL (`http://`, `https://`) o data URL en base64 (`data:image/...`) |
-| Formatos | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Aspect ratio (A / H) | exclusivo `(0.4, 2.5)` |
-| Lado mínimo | ≥ 300 px |
-| Cantidad de imágenes: primer frame I2V | 1 |
-| Cantidad de imágenes: primer + último frame I2V | 2 |
-| **Cantidad de imágenes: R2V (V2 / Fast)** | **1 – 9** |
+### Shared media floors
 
-### Vídeos
-
-| Restricción | Valor |
-|---|---|
-| Métodos de entrada | URL (`http://`, `https://`) o data URL en base64 (`data:video/...`) |
-| Formatos | `.mp4`, `.mov` |
-| Códecs de vídeo | H.264 / AVC, H.265 / HEVC |
-| Códecs de audio (en el contenedor) | AAC, MP3 |
-| **Duración por clip** | `[2, 15]` s (inclusive) |
-| **Cantidad máx. de clips** | 3 (R2V / Stitch / Extend) |
-| **Duración combinada total** | ≤ 15 s entre todos los clips |
-| **Tamaño por clip** | ≤ 50 MB |
-
-### Audio
-
-| Restricción | Valor |
+| Constraint | Value |
 |---|---|
-| Métodos de entrada | URL (`http://`, `https://`) o data URL en base64 (`data:audio/...`) |
-| Formatos | `.wav`, `.mp3` |
-| **Duración por clip** | `[2, 15]` s |
-| **Cantidad máx. de clips** | 3 |
-| **Duración combinada total** | ≤ 15 s entre todos los clips |
-| **Tamaño por clip** | ≤ 15 MB |
-
-El audio de referencia se admite solo en las variantes R2V. Cada entrada se reenvía al modelo como un elemento de contenido `role: "reference_audio"` al que el prompt se dirige como `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — el modelo usa cada clip para timbre vocal, efectos de sonido o música de fondo dependiendo de cómo el prompt lo encuadre. El campo singular heredado `audio_url` mapea a la misma forma de contenido y ahora es equivalente a pasar un `reference_audio_urls` de un elemento.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` no puede ser la única entrada de referencia.** El modelo requiere al menos una referencia de imagen o vídeo junto con cualquier donante de audio. Empareja `reference_audio_urls` con `reference_image_urls`, `reference_video_urls`, `image_url` o `video_url` — los envíos solo de audio se rechazan.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### Tamaño de la solicitud
+### Request size
 
-El endpoint de cola acepta cuerpos JSON de hasta **35 MB**. Las data URLs inline para vídeos grandes pueden empujar más allá de esto — para Stitch multi-clip en particular, prefiere URLs en lugar de base64 inline.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## Precios
+## Pricing
 
-Llama a `POST /api/v1/video/quote` para obtener una cotización para una forma de solicitud dada antes de enviarla a `/video/queue`. El endpoint de cotización es la única fuente autoritativa; los detalles de precio pueden cambiar y no deberían cachearse ni duplicarse en el cliente.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-Cuando los vídeos de referencia forman parte de la solicitud, pasa también `reference_video_total_duration` (la suma de todas las duraciones de clips de referencia en segundos) para que la cotización coincida con lo que cobrará `/video/queue`:
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,9 +228,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Ejemplos completos
+## Complete examples
 
-Todos los ejemplos asumen que `VENICE_API_KEY` está establecida en el entorno.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
 ### Text-to-video
 
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Image-to-video (primer frame)
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video` (y su variante Fast) **no aceptan `aspect_ratio`** — el aspect ratio de salida se deriva automáticamente de las dimensiones de la imagen de entrada. Pasar el campo devuelve un 400 con *"This model does not support aspect_ratio"*. Usa las variantes T2V o R2V si necesitas control explícito de aspect ratio.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Flujo Reference — donante de sujeto
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Flujo Reference — donante de sujeto + audio
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Flujo Edit
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Flujo Edit con anclaje por imagen
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extender hacia adelante
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling de finalización
+### Polling for completion
 
-Tras cada envío a la cola, guarda el `queue_id` devuelto y haz polling de `/video/retrieve` hasta que el cuerpo de la respuesta sea `video/mp4`:
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-La respuesta es JSON (`{ "status": "queued" | "running" | "failed", ... }`) hasta que el trabajo se complete, momento en el que el cuerpo de la respuesta cambia a bytes `video/mp4`. Consulta [Generación de vídeo](/guides/media/video-generation) para el patrón completo de polling.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## Resolución de problemas
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-Los envíos de reference-to-video deben incluir al menos uno de `reference_image_urls`, `reference_video_urls`, `image_references` o `video_references`. La generación pura solo de texto no es un flujo R2V válido — usa `seedance-2-0-text-to-video` en su lugar. `reference_audio_urls` por sí solo no es suficiente (consulta la sección de Audio anterior).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-El modelo limita los vídeos de referencia a 3. Si necesitas más clips, haz un Stitch primero (3 → 1) y luego usa la salida como referencia para un seguimiento.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / agregado `> 15s`
+### Prompt routes to the wrong workflow
 
-La duración por clip es `[2, 15]` segundos **inclusive**; la suma entre todos los vídeos de referencia también está limitada a 15 segundos. Recorta los clips en el cliente antes del envío.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### El prompt enruta al flujo incorrecto
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-El flujo se infiere de la sintaxis del prompt. Errores de enrutamiento comunes:
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- Querer **Extend** pero escribir `Refer to ...` → el modelo trata tu vídeo como *donante*, no como lienzo a continuar
-- Querer **Stitch** pero escribir `Refer to ...` → el modelo elige uno como donante, ignora los demás
-- Querer **Edit** pero escribir `Generate a video based on <Video 1>` → ambiguo; el modelo puede caer en Reference por defecto
+### Face / likeness inputs rejected
 
-Usa los prefijos canónicos exactamente como están escritos: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### La cotización no coincide con el importe en cola
+### Quote doesn't match the queued amount
 
-Si incluiste un vídeo de referencia pero no pasaste `reference_video_total_duration` a `/video/quote`, la cotización y el importe en cola pueden diferir. Pasa siempre `reference_video_total_duration` (suma de todas las duraciones de clips de referencia, en segundos) cuando haya vídeos de referencia.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## Referencias
+## References
 
-- Endpoint de cola de vídeo de Venice: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Endpoint de cotización de Venice: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Guía complementaria: [Reference to Video](/guides/media/reference-to-video) (cubre Kling O3 + Grok Imagine R2V)
-- Guía complementaria: [Generación de vídeo](/guides/media/video-generation) (visión general de cola / polling)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/fr/api-reference/endpoint/video/queue.mdx b/fr/api-reference/endpoint/video/queue.mdx
index e6902bb8..5c21a917 100644
--- a/fr/api-reference/endpoint/video/queue.mdx
+++ b/fr/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Appelez `/video/quote` pour obtenir une estimation du prix, puis interrogez `/vi
 
 Les modèles privés renvoient également un `download_url` pour la vidéo terminée. Il s'agit d'une URL de livraison à durée de vie courte (quelques relances sont acceptables si un téléchargement échoue) ; voir le [guide de génération vidéo](/guides/media/video-generation#private-download-links) pour les détails et la méthode `DELETE` optionnelle pour la confidentialité.
 
-### Seedance 2.0
+### Seedance 2.0 et 2.5
 
-Pour les modèles `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*`), voir le [guide Seedance 2.0](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les modèles de prompt canoniques, les limites d'entrée multimodale et les détails de tarification.
+Pour les modèles `seedance-2-0-*` et `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*` de Seedance 2.0), voir le [Guide Seedance 2.0 et 2.5](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les limites multimodales par famille, le comportement faces-off de l’API et les tarifs.
 
-### Consentement Seedance
+### API Seedance faces-off
 
-Lorsqu'une requête Seedance image-to-video ou reference-to-video contient un visage humain, l'API renvoie un `409 needs_consent` avec le texte de la politique, et vous resoumettez la même requête avec une attestation `consents.seedance`. Voir le [guide de consentement pour les médias avec visage Seedance](/guides/media/seedance-face-consent) pour le flux complet, l'objet de consentement, la déduplication et la révocation.
+Les modèles Seedance publics **ne** prennent **pas** en charge le provisioning d’assets de visage ni l’attestation de consentement. Les entrées comportant un visage peuvent être rejetées en amont. Pour la génération likeness / face assets, utilisez l’app ou Studio Venice. Voir le [guide Seedance](/guides/media/seedance-2-0).
 
 ### Agrandissement vidéo
 
diff --git a/fr/guides/media/seedance-2-0.mdx b/fr/guides/media/seedance-2-0.mdx
index ace35ba7..91e41d03 100644
--- a/fr/guides/media/seedance-2-0.mdx
+++ b/fr/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "Générez, éditez, prolongez et assemblez des vidéos avec Seedance 2.0 sur Venice : workflows text, image et reference-to-video sur un seul endpoint."
-'og:title': "Guide Seedance 2.0 | Documentation de l'API Venice"
-'og:description': "Générez, éditez, prolongez et assemblez des vidéos avec Seedance 2.0 via l'API Venice. Modèles de workflow, limites multimodales, tarification et exemples concrets."
+title: "Seedance 2.0 et 2.5"
+description: "Générez, modifiez, prolongez et assemblez des vidéos avec Seedance 2.0 et 2.5 sur Venice - workflows texte/image/référence, API faces-off et limites multimodales par famille."
+'og:title': "Guide Seedance 2.0 et 2.5 | Venice API Docs"
+'og:description': "Générez, modifiez, prolongez et assemblez des vidéos avec Seedance 2.0 et 2.5 via l'API Venice. Workflows, faces-off, limites, tarifs et exemples."
 ---
 
-Seedance 2.0 est un modèle vidéo multimodal phare exposé sur Venice sous forme d'une famille de trois variantes pour la génération vidéo pilotée par texte, image et référence. La variante **reference-to-video** est exceptionnellement puissante : un seul endpoint et un seul identifiant de modèle gèrent **quatre workflows distincts** (Reference, Edit, Extend, Stitch) — le workflow est inféré à partir de la **forme de votre prompt**.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-Ce guide passe en revue les variantes, les quatre workflows avec leurs prompts canoniques, les limites d'entrée multimodale, la tarification et des exemples `curl` complets.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## Variantes
+<Warning>
+**Les visages ne sont pas pris en charge sur l'API Seedance publique.** Venice n'exécute pas le provisioning d'assets de visage ni l'attestation de consentement (`consents.seedance` / `needs_consent`) pour les modèles Seedance publics. Les entrées comportant un visage peuvent être rejetées en amont (politique de contenu / fournisseur). Pour la génération likeness / face assets, utilisez l'app ou Studio Venice.
+</Warning>
+
+## Variants
 
-| Identifiant du modèle | Variante | Résolutions de sortie | Notes |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Prompt textuel uniquement |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Ancrage par première image (et optionnellement dernière image) |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Jusqu'à 9 images de référence + 3 vidéos de référence + 3 audios de référence. Alimente Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Palier plus rapide, fidélité moindre |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Palier plus rapide, fidélité moindre |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Palier plus rapide, fidélité moindre ; même ensemble de workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-Toutes les variantes sont asynchrones. Soumettez via `POST /api/v1/video/queue`, puis interrogez `POST /api/v1/video/retrieve` jusqu'à ce que le corps de la réponse soit `video/mp4`. Voir [Génération vidéo](/guides/media/video-generation) pour le flux général de file d'attente.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## Le modèle « un modèle, quatre workflows »
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-La variante reference-to-video (`seedance-2-0-reference-to-video` et son équivalent Fast) est le même modèle sous-jacent servant quatre tâches différentes. **Le modèle déduit la tâche à partir du préfixe du prompt et de la forme de vos entrées.** Il n'y a pas de champ `task` ou `workflow` — la syntaxe du prompt est le routage.
+## The "one model, four workflows" model
 
-| Workflow | Ce qu'il fait | Préfixe de prompt | Entrées |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Génère une nouvelle vidéo en utilisant les fichiers de référence téléversés comme donneurs pour le sujet / le mouvement / le style / l'audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Texte + ≥1 référence image OU vidéo (0-9 images, 0-3 vidéos), plus optionnellement jusqu'à 3 donneurs audio |
-| **Edit** | Modifie une seule vidéo d'entrée tout en préservant le reste | `Strictly edit <Video 1>, changing its ...` | 1 vidéo d'entrée + texte (ancrage par images optionnel) |
-| **Extend** | Prolongation avant / arrière d'un clip | `Extend <Video 1>, generate ...` | 1 vidéo d'entrée + texte |
-| **Stitch** | Assemble 2-3 clips avec des transitions générées automatiquement | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 vidéos d'entrée + texte |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-La **syntaxe du prompt est canonique et sensible à la casse** : crochets pointus, première lettre en majuscule, espace unique avant le numéro — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Modèles de workflow
+## Workflow patterns
 
-### Workflow Reference
+### Reference workflow
 
-Utilise les fichiers de référence téléversés comme **donneurs** — sujet, scène, mouvement, style, timbre vocal — pour générer une toute nouvelle vidéo.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**Modèles de prompt canoniques** :
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,23 +60,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Exemples** :
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (les donneurs audio doivent être associés à au moins une référence image ou vidéo — l'audio seul est rejeté)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Workflow Edit
+### Edit workflow
 
-Modifie une seule vidéo d'entrée. **Tout ce qui n'est pas explicitement nommé dans le prompt est préservé.** Utilisez-le lorsque vous voulez un changement localisé (remplacement de sujet, changement de météo/couleur, ajout/suppression d'élément) plutôt qu'une vidéo entièrement nouvelle.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**Modèle de prompt canonique** :
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sous-modèles pour un contrôle plus fin** :
+**Sub-patterns for finer control**:
 
 ```
 Add Elements:
@@ -80,57 +89,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Exemples** :
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-Le dernier exemple combine Edit avec une référence d'image — parfaitement légal, le modèle utilise `<Image 1>` comme donneur visuel pour le remplacement.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Workflow Extend
+### Extend workflow
 
-Prolonge un seul clip vers l'avant ou l'arrière dans le temps. **Par défaut, Seedance ne renvoie que le nouveau contenu** — pas l'entrée originale concaténée avec l'extension. C'est intentionnel, pour la continuité de transition ; si vous voulez que le clip d'entrée soit préservé à côté de l'extension, indiquez-le explicitement :
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← préserve l'entrée au début
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← préserve l'entrée à la fin
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-Gestion des transitions : le modèle extrait automatiquement les images de transition pour un fondu fluide, et les segments originaux de la vidéo d'entrée ne sont pas régénérés.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**Exemples** :
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Workflow Stitch (Track Completion)
+### Stitch workflow (Track Completion)
 
-Connecte 2-3 clips d'entrée avec des transitions générées par IA. La durée totale d'entrée combinée doit être ≤ 15 s.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**Modèle de prompt canonique** :
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Exemples** :
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-Le modèle ajuste automatiquement les segments de liaison aux points de jonction pour assurer la continuité.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## Formule universelle de prompt
+## Universal prompt formula
 
-À travers les quatre workflows, la formule de rédaction recommandée est :
+Across all four workflows, the recommended authoring formula is:
 
 ```
 Subject + Motion + Environment (Optional)
@@ -139,72 +148,70 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion** : la base logique — définit « qui » effectue « quelle action »
-- **Environment + Aesthetics** : arrière-plan spatial, éclairage, style visuel
-- **Camera** : type ou mouvement de prise de vue explicite
-- **Audio** : effets sonores ambiants ou direction vocale pour une sortie immersive
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-Superposer cela à un préfixe de workflow (par exemple, `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produit les sorties de plus haute qualité.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## Limites d'entrée multimodale
+## Multimodal input limits
 
-Les valeurs ci-dessous sont celles que l'API Venice accepte. Les requêtes hors de ces plages sont rejetées au niveau du schéma avec un 400 avant d'atteindre l'inférence.
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### Images
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| Contrainte | Valeur |
-|---|---|
-| Méthodes d'entrée | URL (`http://`, `https://`) ou URL data Base64 (`data:image/...`) |
-| Formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Ratio d'aspect (L / H) | exclusif `(0.4, 2.5)` |
-| Côté minimum | ≥ 300 px |
-| Nombre d'images : première image I2V | 1 |
-| Nombre d'images : première + dernière image I2V | 2 |
-| **Nombre d'images : R2V (V2 / Fast)** | **1 – 9** |
+### Shared media floors
 
-### Vidéos
-
-| Contrainte | Valeur |
-|---|---|
-| Méthodes d'entrée | URL (`http://`, `https://`) ou URL data Base64 (`data:video/...`) |
-| Formats | `.mp4`, `.mov` |
-| Codecs vidéo | H.264 / AVC, H.265 / HEVC |
-| Codecs audio (dans le conteneur) | AAC, MP3 |
-| **Durée par clip** | `[2, 15]` s (inclusif) |
-| **Nombre maximum de clips** | 3 (R2V / Stitch / Extend) |
-| **Durée totale combinée** | ≤ 15 s à travers tous les clips |
-| **Taille par clip** | ≤ 50 Mo |
-
-### Audio
-
-| Contrainte | Valeur |
+| Constraint | Value |
 |---|---|
-| Méthodes d'entrée | URL (`http://`, `https://`) ou URL data Base64 (`data:audio/...`) |
-| Formats | `.wav`, `.mp3` |
-| **Durée par clip** | `[2, 15]` s |
-| **Nombre maximum de clips** | 3 |
-| **Durée totale combinée** | ≤ 15 s à travers tous les clips |
-| **Taille par clip** | ≤ 15 Mo |
-
-L'audio de référence n'est pris en charge que sur les variantes R2V. Chaque entrée est transmise au modèle en tant qu'élément de contenu `role: "reference_audio"` que le prompt adresse comme `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — le modèle utilise chaque clip pour le timbre vocal, les effets sonores ou la musique de fond selon la façon dont le prompt le formule. L'ancien champ singulier `audio_url` mappe vers la même forme de contenu et équivaut désormais au passage d'un `reference_audio_urls` à un seul élément.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` ne peut pas être la seule entrée de référence.** Le modèle requiert au moins une référence image ou vidéo en plus de tout donneur audio. Associez `reference_audio_urls` à `reference_image_urls`, `reference_video_urls`, `image_url` ou `video_url` — les soumissions audio seules sont rejetées.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### Taille de la requête
+### Request size
 
-L'endpoint de file d'attente accepte des corps JSON jusqu'à **35 Mo**. Les URL data en ligne pour de grandes vidéos peuvent dépasser cette limite — pour le Stitch multi-clips en particulier, préférez les URL aux base64 en ligne.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## Tarification
+## Pricing
 
-Appelez `POST /api/v1/video/quote` pour obtenir un devis pour une forme de requête donnée avant de la soumettre à `/video/queue`. L'endpoint de devis est la seule source faisant autorité ; les détails de tarification peuvent changer et ne doivent pas être mis en cache ou dupliqués côté client.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-Lorsque des vidéos de référence font partie de la requête, transmettez également `reference_video_total_duration` (la somme de toutes les durées des clips de référence en secondes) afin que le devis corresponde à ce que `/video/queue` facturera :
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,9 +228,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Exemples complets
+## Complete examples
 
-Tous les exemples supposent que `VENICE_API_KEY` est défini dans l'environnement.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
 ### Text-to-video
 
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Image-to-video (première image)
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video` (et sa variante Fast) **n'accepte pas `aspect_ratio`** — le ratio d'aspect de sortie est dérivé automatiquement des dimensions de l'image d'entrée. Passer ce champ renvoie un 400 avec *« This model does not support aspect_ratio »*. Utilisez les variantes T2V ou R2V si vous avez besoin d'un contrôle explicite du ratio d'aspect.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Workflow Reference — donneur de sujet
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Reference — donneur sujet + audio
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Edit
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Edit avec ancrage par image
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend vers l'avant
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Sondage jusqu'à la fin
+### Polling for completion
 
-Après chaque soumission de file d'attente, enregistrez le `queue_id` retourné et interrogez `/video/retrieve` jusqu'à ce que le corps de la réponse soit `video/mp4` :
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-La réponse est en JSON (`{ "status": "queued" | "running" | "failed", ... }`) jusqu'à ce que la tâche se termine, moment auquel le corps de la réponse bascule vers des octets `video/mp4`. Voir [Génération vidéo](/guides/media/video-generation) pour le modèle complet de sondage.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## Dépannage
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-Les soumissions reference-to-video doivent inclure au moins l'un de `reference_image_urls`, `reference_video_urls`, `image_references` ou `video_references`. Une génération purement textuelle n'est pas un workflow R2V valide — utilisez plutôt `seedance-2-0-text-to-video`. `reference_audio_urls` seul ne suffit pas (voir la section Audio ci-dessus).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-Le modèle plafonne les vidéos de référence à 3. Si vous avez besoin de plus de clips, exécutez d'abord un Stitch (3 → 1), puis utilisez la sortie comme référence pour une suite.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / agrégat `> 15s`
+### Prompt routes to the wrong workflow
 
-La durée par clip est `[2, 15]` secondes **inclusif** ; la somme à travers toutes les vidéos de référence est également plafonnée à 15 secondes. Coupez les clips côté client avant la soumission.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### Le prompt achemine vers le mauvais workflow
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-Le workflow est inféré à partir de la syntaxe du prompt. Erreurs d'acheminement courantes :
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- Vouloir **Extend** mais écrire `Refer to ...` → le modèle traite votre vidéo comme un *donneur*, pas comme une toile à continuer
-- Vouloir **Stitch** mais écrire `Refer to ...` → le modèle en choisit une comme donneur, ignore les autres
-- Vouloir **Edit** mais écrire `Generate a video based on <Video 1>` → ambigu ; le modèle peut basculer par défaut sur Reference
+### Face / likeness inputs rejected
 
-Utilisez les préfixes canoniques exactement tels qu'écrits : `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### Le devis ne correspond pas au montant en file d'attente
+### Quote doesn't match the queued amount
 
-Si vous avez inclus une vidéo de référence mais n'avez pas transmis `reference_video_total_duration` à `/video/quote`, le devis et le montant en file d'attente peuvent différer. Transmettez toujours `reference_video_total_duration` (somme de toutes les durées des clips de référence, en secondes) lorsque des vidéos de référence sont présentes.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## Références
+## References
 
-- Endpoint de file d'attente vidéo Venice : [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Endpoint de devis Venice : [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Guide complémentaire : [Reference to Video](/guides/media/reference-to-video) (couvre Kling O3 + Grok Imagine R2V)
-- Guide complémentaire : [Génération vidéo](/guides/media/video-generation) (vue d'ensemble queue / polling)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/guides/media/seedance-2-0.mdx b/guides/media/seedance-2-0.mdx
index 50014f43..d53ac7cb 100644
--- a/guides/media/seedance-2-0.mdx
+++ b/guides/media/seedance-2-0.mdx
@@ -1,13 +1,17 @@
 ---
-title: "Seedance 2.0"
-description: "Generate, edit, extend, and stitch videos with Seedance 2.0 on Venice — text, image, and reference-to-video workflows on one endpoint, with worked examples."
-'og:title': "Seedance 2.0 Guide | Venice API Docs"
-'og:description': "Generate, edit, extend and stitch videos with Seedance 2.0 via the Venice API. Workflow patterns, multimodal limits, pricing, and worked examples."
+title: "Seedance 2.0 & 2.5"
+description: "Generate, edit, extend, and stitch videos with Seedance 2.0 and 2.5 on Venice - text, image, and reference-to-video workflows, faces-off API behavior, and family-specific multimodal limits."
+'og:title': "Seedance 2.0 & 2.5 Guide | Venice API Docs"
+'og:description': "Generate, edit, extend and stitch videos with Seedance 2.0 and 2.5 via the Venice API. Workflow patterns, faces-off behavior, multimodal limits, pricing, and worked examples."
 ---
 
-Seedance 2.0 is a flagship multimodal video model exposed on Venice as a family of three variants for text-, image-, and reference-driven video generation. The **reference-to-video** variant is unusually powerful: a single endpoint and a single model ID handle **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-This guide walks through the variants, the four workflows with their canonical prompts, the multimodal input limits, pricing, and complete `curl` examples.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+
+<Warning>
+**Faces are not supported on the public Seedance API.** Venice does not run face-asset provisioning or face-consent attestation (`consents.seedance` / `needs_consent`) for public Seedance models. Face-bearing inputs may be rejected upstream as a content-policy / provider error. For likeness / face-asset generation, use the Venice app or Studio.
+</Warning>
 
 ## Variants
 
@@ -19,19 +23,24 @@ This guide walks through the variants, the four workflows with their canonical p
 | `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
 | `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
 | `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+
 ## The "one model, four workflows" model
 
-The reference-to-video variant (`seedance-2-0-reference-to-video` and its Fast sibling) is the same underlying model serving four different tasks. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference (0-9 images, 0-3 videos), plus optionally up to 3 audio donors |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
 | **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
 | **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch 2-3 clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 input videos + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
 The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
@@ -110,7 +119,7 @@ Transition handling: the model automatically extracts the transition frames for
 
 ### Stitch workflow (Track Completion)
 
-Connect 2-3 input clips with AI-generated transitions. Total combined input duration must be ≤ 15 s.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
 **Canonical prompt pattern**:
 
@@ -152,43 +161,41 @@ Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, chang
 
 Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### Images
-
-| Constraint | Value |
-|---|---|
-| Input methods | URL (`http://`, `https://`) or Base64 data URL (`data:image/...`) |
-| Formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Minimum side | ≥ 300 px |
-| Image count: I2V first-frame | 1 |
-| Image count: I2V first + last frame | 2 |
-| **Image count: R2V (V2 / Fast)** | **1 – 9** |
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-### Videos
+### Shared media floors
 
 | Constraint | Value |
 |---|---|
-| Input methods | URL (`http://`, `https://`) or Base64 data URL (`data:video/...`) |
-| Formats | `.mp4`, `.mov` |
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
 | Video codecs | H.264 / AVC, H.265 / HEVC |
 | Audio codecs (in container) | AAC, MP3 |
-| **Duration per clip** | `[2, 15]` s (inclusive) |
-| **Max clip count** | 3 (R2V / Stitch / Extend) |
-| **Total combined duration** | ≤ 15 s across all clips |
-| **Per-clip size** | ≤ 50 MB |
-
-### Audio
-
-| Constraint | Value |
-|---|---|
-| Input methods | URL (`http://`, `https://`) or Base64 data URL (`data:audio/...`) |
-| Formats | `.wav`, `.mp3` |
-| **Duration per clip** | `[2, 15]` s |
-| **Max clip count** | 3 |
-| **Total combined duration** | ≤ 15 s across all clips |
-| **Per-clip size** | ≤ 15 MB |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
 **`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
@@ -240,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
 ### Image-to-video (first frame)
 
 ```bash
@@ -256,7 +278,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video` (and its Fast variant) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -275,6 +297,25 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
 ### Reference workflow — subject + audio donor
 
 ```bash
@@ -388,15 +429,18 @@ The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) unti
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use `seedance-2-0-text-to-video` instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
 
-### `reference_video_urls must have at most 3 videos`
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
 
-The model caps reference videos at 3. If you need more clips, run a Stitch first (3 → 1), then use the output as a reference for a follow-up.
+### Duration / aggregate duration errors
 
-### `Per clip must be 2–15s` / aggregate `> 15s`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-Per-clip duration is `[2, 15]` seconds **inclusive**; the sum across all reference videos is also capped at 15 seconds. Trim clips client-side before submission.
+Trim clips client-side before submission.
 
 ### Prompt routes to the wrong workflow
 
@@ -408,6 +452,10 @@ Workflow is inferred from prompt syntax. Common misroutings:
 
 Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
+### Face / likeness inputs rejected
+
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+
 ### Quote doesn't match the queued amount
 
 If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
diff --git a/it/api-reference/endpoint/video/queue.mdx b/it/api-reference/endpoint/video/queue.mdx
index c65b84ad..9ae50b97 100644
--- a/it/api-reference/endpoint/video/queue.mdx
+++ b/it/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Chiama `/video/quote` per ottenere una stima del prezzo, quindi esegui il pollin
 
 I modelli privati restituiscono anche un `download_url` per il video finito. Si tratta di un URL di consegna a breve termine (alcuni retry vanno bene se il download cade); consulta la [guida alla generazione video](/guides/media/video-generation#private-download-links) per i dettagli e l'uso opzionale di `DELETE` per la privacy.
 
-### Seedance 2.0
+### Seedance 2.0 e 2.5
 
-Per i modelli `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*`), consulta la [guida a Seedance 2.0](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i pattern di prompt canonici, i limiti di input multimodali e i dettagli sui prezzi.
+Per i modelli `seedance-2-0-*` e `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*` di Seedance 2.0), vedi la [Guida Seedance 2.0 e 2.5](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i limiti multimodali per famiglia, il comportamento faces-off dell’API e i dettagli sui prezzi.
 
-### Consenso Seedance
+### API Seedance faces-off
 
-Quando una richiesta image- o reference-to-video di Seedance contiene un volto umano, l'API restituisce un `409 needs_consent` con il testo della policy e tu reinvii la stessa richiesta con un'attestazione `consents.seedance`. Consulta la [guida al consenso per i contenuti facciali di Seedance](/guides/media/seedance-face-consent) per il flusso completo, l'oggetto di consenso, la deduplica e la revoca.
+I modelli Seedance pubblici **non** supportano il provisioning di face asset né l’attestazione di consenso sul volto. Gli input con volti possono essere rifiutati upstream. Per la generazione likeness / face asset usa l’app o Studio Venice. Vedi la [guida Seedance](/guides/media/seedance-2-0).
 
 ### Upscaling video
 
diff --git a/it/guides/media/seedance-2-0.mdx b/it/guides/media/seedance-2-0.mdx
index 188a9c4d..c9f977db 100644
--- a/it/guides/media/seedance-2-0.mdx
+++ b/it/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "Genera, modifica, estendi e cuci video con Seedance 2.0 su Venice — workflow text-, image- e reference-to-video su un unico endpoint, con esempi pratici."
-'og:title': "Seedance 2.0 Guide | Venice API Docs"
-'og:description': "Genera, modifica, estendi e cuci video con Seedance 2.0 tramite l'API Venice. Pattern di workflow, limiti multimodali, prezzi ed esempi completi."
+title: "Seedance 2.0 e 2.5"
+description: "Genera, modifica, estendi e unisci video con Seedance 2.0 e 2.5 su Venice - workflow testo/immagine/riferimento, API faces-off e limiti multimodali per famiglia."
+'og:title': "Guida Seedance 2.0 e 2.5 | Venice API Docs"
+'og:description': "Genera, modifica, estendi e unisci video con Seedance 2.0 e 2.5 tramite l'API Venice. Workflow, faces-off, limiti, prezzi ed esempi."
 ---
 
-Seedance 2.0 è un modello video multimodale flagship esposto su Venice come famiglia di tre varianti per la generazione video text-, image- e reference-driven. La variante **reference-to-video** è straordinariamente potente: un singolo endpoint e un singolo ID di modello gestiscono **quattro workflow distinti** (Reference, Edit, Extend, Stitch) — il workflow viene dedotto dalla **forma del tuo prompt**.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-Questa guida illustra le varianti, i quattro workflow con i loro prompt canonici, i limiti di input multimodali, i prezzi e esempi `curl` completi.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## Varianti
+<Warning>
+**I volti non sono supportati sull'API Seedance pubblica.** Venice non esegue provisioning di face asset né attestazione di consenso (`consents.seedance` / `needs_consent`) per i modelli Seedance pubblici. Gli input con volti possono essere rifiutati upstream come errore di content policy / provider. Per la generazione likeness / face asset usa l'app o Studio Venice.
+</Warning>
+
+## Variants
 
-| Model ID | Variante | Risoluzioni di output | Note |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Solo prompt testuale |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Grounding su first-frame (e opzionalmente last-frame) |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Fino a 9 immagini di riferimento + 3 video di riferimento + 3 donor audio di riferimento. Alimenta Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Tier più veloce, fidelity inferiore |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Tier più veloce, fidelity inferiore |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Tier più veloce, fidelity inferiore; stesso set di workflow |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-Tutte le varianti sono asincrone. Invia tramite `POST /api/v1/video/queue`, poi fai polling su `POST /api/v1/video/retrieve` finché il corpo della risposta non è `video/mp4`. Vedi [Video Generation](/guides/media/video-generation) per il flusso generale di queue.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## Il modello "one model, four workflows"
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-La variante reference-to-video (`seedance-2-0-reference-to-video` e il suo fratello Fast) è lo stesso modello sottostante che serve quattro task diversi. **Il modello deduce il task dal prefisso del prompt e dalla forma dei tuoi input.** Non c'è un campo `task` o `workflow` — la sintassi del prompt è il routing.
+## The "one model, four workflows" model
 
-| Workflow | Cosa fa | Prefisso del prompt | Input |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Genera un nuovo video usando i file di riferimento caricati come donor per soggetto / movimento / stile / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Testo + ≥1 immagine OR video di riferimento (0-9 immagini, 0-3 video), più opzionalmente fino a 3 donor audio |
-| **Edit** | Modifica un singolo video di input preservando il resto | `Strictly edit <Video 1>, changing its ...` | 1 video di input + testo (immagini opzionali per grounding) |
-| **Extend** | Estensione in avanti / all'indietro di una clip | `Extend <Video 1>, generate ...` | 1 video di input + testo |
-| **Stitch** | Unisci 2-3 clip con transizioni auto-generate | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 video di input + testo |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-La **sintassi del prompt è canonica e case-sensitive**: parentesi angolari, prima lettera maiuscola, singolo spazio prima del numero — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Pattern di workflow
+## Workflow patterns
 
-### Workflow Reference
+### Reference workflow
 
-Usa i file di riferimento caricati come **donor** — soggetto, scena, movimento, stile, timbro vocale — per generare un video completamente nuovo.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**Pattern canonici del prompt**:
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,160 +60,158 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Esempi**:
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (i donor audio devono essere abbinati ad almeno un'immagine o video di riferimento — solo audio viene rifiutato)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Workflow Edit
+### Edit workflow
 
-Modifica un singolo video di input. **Tutto ciò che non è esplicitamente nominato nel prompt viene preservato.** Usalo quando vuoi un cambiamento localizzato (sostituzione del soggetto, cambio di meteo/colore, aggiunta/rimozione di elementi) piuttosto che un video completamente nuovo.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**Pattern canonico del prompt**:
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sotto-pattern per un controllo più fine**:
+**Sub-patterns for finer control**:
 
 ```
-Aggiungere elementi:
+Add Elements:
   At [timestamp / timing] and [spatial location] of <Video 1>, add [description of intended element].
 
-Rimuovere elementi:
+Remove Elements:
   Remove [element to be deleted] from <Video 1>, keeping the rest of the video content unchanged.
 
-Modificare elementi:
+Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Esempi**:
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-L'ultimo esempio combina Edit con un riferimento di immagine — è perfettamente legale, il modello usa `<Image 1>` come donor visivo per la sostituzione.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Workflow Extend
+### Extend workflow
 
-Continua una singola clip in avanti o all'indietro nel tempo. **Di default Seedance restituisce solo il nuovo contenuto** — non l'input originale concatenato con l'estensione. Questo è di proposito, per la continuità della transizione; se vuoi che la clip di input sia preservata accanto all'estensione, dillo esplicitamente:
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserva l'input all'inizio
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserva l'input alla fine
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-Gestione della transizione: il modello estrae automaticamente i frame di transizione per un blending senza soluzione di continuità, e i segmenti originali del video di input non vengono ri-generati.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**Esempi**:
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Workflow Stitch (Track Completion)
+### Stitch workflow (Track Completion)
 
-Collega 2-3 clip di input con transizioni generate dall'AI. La durata combinata totale di input deve essere ≤ 15 s.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**Pattern canonico del prompt**:
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Esempi**:
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-Il modello taglia automaticamente i segmenti di connessione nei punti di giuntura per garantire continuità.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## Formula universale del prompt
+## Universal prompt formula
 
-In tutti e quattro i workflow, la formula di scrittura consigliata è:
+Across all four workflows, the recommended authoring formula is:
 
 ```
-Subject + Motion + Environment (Opzionale)
-       + Camera Movement / Cut (Opzionale)
-       + Aesthetic Description (Opzionale)
-       + Audio (Opzionale)
+Subject + Motion + Environment (Optional)
+       + Camera Movement / Cut (Optional)
+       + Aesthetic Description (Optional)
+       + Audio (Optional)
 ```
 
-- **Subject + Motion**: la base logica — definisce "Chi" sta eseguendo "Quale azione"
-- **Environment + Aesthetics**: sfondo spaziale, illuminazione, stile visivo
-- **Camera**: tipo di inquadratura o movimento esplicito
-- **Audio**: effetti sonori ambientali o direzione vocale per un output immersivo
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-Sovrapporre questo a un prefisso di workflow (es. `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produce gli output di qualità più alta.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## Limiti di input multimodale
+## Multimodal input limits
 
-I valori sottostanti sono quelli che l'API Venice accetta. Le richieste al di fuori di questi range vengono rifiutate a livello di schema con un 400 prima di raggiungere l'inferenza.
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### Immagini
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| Vincolo | Valore |
-|---|---|
-| Metodi di input | URL (`http://`, `https://`) o data URL Base64 (`data:image/...`) |
-| Formati | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Aspect ratio (W / H) | esclusivo `(0.4, 2.5)` |
-| Lato minimo | ≥ 300 px |
-| Conteggio immagini: I2V first-frame | 1 |
-| Conteggio immagini: I2V first + last frame | 2 |
-| **Conteggio immagini: R2V (V2 / Fast)** | **1 – 9** |
+### Shared media floors
 
-### Video
-
-| Vincolo | Valore |
-|---|---|
-| Metodi di input | URL (`http://`, `https://`) o data URL Base64 (`data:video/...`) |
-| Formati | `.mp4`, `.mov` |
-| Codec video | H.264 / AVC, H.265 / HEVC |
-| Codec audio (nel container) | AAC, MP3 |
-| **Durata per clip** | `[2, 15]` s (inclusivo) |
-| **Numero massimo di clip** | 3 (R2V / Stitch / Extend) |
-| **Durata totale combinata** | ≤ 15 s tra tutte le clip |
-| **Dimensione per clip** | ≤ 50 MB |
-
-### Audio
-
-| Vincolo | Valore |
+| Constraint | Value |
 |---|---|
-| Metodi di input | URL (`http://`, `https://`) o data URL Base64 (`data:audio/...`) |
-| Formati | `.wav`, `.mp3` |
-| **Durata per clip** | `[2, 15]` s |
-| **Numero massimo di clip** | 3 |
-| **Durata totale combinata** | ≤ 15 s tra tutte le clip |
-| **Dimensione per clip** | ≤ 15 MB |
-
-L'audio di riferimento è supportato solo sulle varianti R2V. Ogni entry viene inoltrata al modello come content item `role: "reference_audio"` che il prompt richiama come `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — il modello usa ciascuna clip per timbro vocale, effetti sonori o musica di sottofondo a seconda di come la inquadra il prompt. Il campo legacy singolare `audio_url` viene mappato sulla stessa forma di contenuto ed è ora equivalente a passare un `reference_audio_urls` di un solo elemento.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` non può essere l'unico input di riferimento.** Il modello richiede almeno un'immagine o un video di riferimento accanto a qualsiasi donor audio. Abbina `reference_audio_urls` con `reference_image_urls`, `reference_video_urls`, `image_url` o `video_url` — le submission solo audio vengono rifiutate.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### Dimensione della richiesta
+### Request size
 
-L'endpoint di queue accetta corpi JSON fino a **35 MB**. I data URL inline per video di grandi dimensioni possono superare questa soglia — per Stitch multi-clip in particolare, preferisci gli URL al base64 inline.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## Prezzi
+## Pricing
 
-Chiama `POST /api/v1/video/quote` per ottenere un preventivo per una determinata forma di richiesta prima di inviarla a `/video/queue`. L'endpoint quote è l'unica fonte autorevole; i dettagli sui prezzi possono cambiare e non dovrebbero essere messi in cache o duplicati lato client.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-Quando uno o più video di riferimento fanno parte della richiesta, passa anche `reference_video_total_duration` (la somma di tutte le durate delle clip di riferimento in secondi) in modo che il preventivo corrisponda a quello che `/video/queue` addebiterà:
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,9 +228,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Esempi completi
+## Complete examples
 
-Tutti gli esempi assumono che `VENICE_API_KEY` sia impostata nell'environment.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
 ### Text-to-video
 
@@ -240,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
 ### Image-to-video (first frame)
 
 ```bash
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video` (e la sua variante Fast) **non accettano `aspect_ratio`** — l'aspect ratio di output viene auto-derivato dalle dimensioni dell'immagine di input. Passare il campo restituisce un 400 con *"This model does not support aspect_ratio"*. Usa le varianti T2V o R2V se hai bisogno di un controllo esplicito sull'aspect ratio.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Workflow Reference — subject donor
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Reference — subject + audio donor
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Edit
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Edit con image grounding
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend in avanti
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -344,7 +385,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clip)
+### Stitch (3 clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling per il completamento
+### Polling for completion
 
-Dopo ogni submission in queue, salva il `queue_id` restituito e fai polling su `/video/retrieve` finché il corpo della risposta non è `video/mp4`:
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,7 +421,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-La risposta è JSON (`{ "status": "queued" | "running" | "failed", ... }`) finché il job non è completato, momento in cui il corpo della risposta passa a byte `video/mp4`. Vedi [Video Generation](/guides/media/video-generation) per il pattern di polling completo.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
@@ -388,35 +429,42 @@ La risposta è JSON (`{ "status": "queued" | "running" | "failed", ... }`) finch
 
 ### `At least one reference is required for this model`
 
-Le submission reference-to-video devono includere almeno uno tra `reference_image_urls`, `reference_video_urls`, `image_references` o `video_references`. La generazione puramente solo-testo non è un workflow R2V valido — usa invece `seedance-2-0-text-to-video`. `reference_audio_urls` da solo non è sufficiente (vedi la sezione Audio sopra).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-Il modello limita i video di riferimento a 3. Se hai bisogno di più clip, esegui prima uno Stitch (3 → 1), poi usa l'output come riferimento per un follow-up.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / aggregato `> 15s`
+### Prompt routes to the wrong workflow
 
-La durata per clip è `[2, 15]` secondi **inclusivo**; la somma su tutti i video di riferimento è anche limitata a 15 secondi. Taglia le clip lato client prima della submission.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### Il prompt viene instradato al workflow sbagliato
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-Il workflow è dedotto dalla sintassi del prompt. Misroute comuni:
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- Volere **Extend** ma scrivere `Refer to ...` → il modello tratta il tuo video come un *donor*, non come una canvas da continuare
-- Volere **Stitch** ma scrivere `Refer to ...` → il modello sceglie uno come donor e ignora gli altri
-- Volere **Edit** ma scrivere `Generate a video based on <Video 1>` → ambiguo; il modello potrebbe ripiegare su Reference
+### Face / likeness inputs rejected
 
-Usa i prefissi canonici esattamente come scritti: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### Il quote non corrisponde all'importo della queue
+### Quote doesn't match the queued amount
 
-Se hai incluso un video di riferimento ma non hai passato `reference_video_total_duration` a `/video/quote`, il quote e l'importo della queue possono differire. Passa sempre `reference_video_total_duration` (somma di tutte le durate delle clip di riferimento, in secondi) quando sono presenti video di riferimento.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## Riferimenti
+## References
 
-- Endpoint Venice video queue: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Endpoint Venice quote: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Guida companion: [Reference to Video](/guides/media/reference-to-video) (copre Kling O3 + Grok Imagine R2V)
-- Guida companion: [Video Generation](/guides/media/video-generation) (panoramica queue / polling)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/ko/api-reference/endpoint/video/queue.mdx b/ko/api-reference/endpoint/video/queue.mdx
index 9ba663af..c90ddc41 100644
--- a/ko/api-reference/endpoint/video/queue.mdx
+++ b/ko/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ openapi: 'POST /video/queue'
 
 비공개 모델은 완성된 비디오를 위한 `download_url`도 반환합니다. 이는 수명이 짧은 전달 URL입니다(다운로드가 끊겨도 몇 번의 재시도는 괜찮습니다). 자세한 내용과 개인정보 보호를 위한 선택적 `DELETE`는 [비디오 생성 가이드](/guides/media/video-generation#private-download-links)를 참조하세요.
 
-### Seedance 2.0
+### Seedance 2.0 & 2.5
 
-`seedance-2-0-*` 모델(text-to-video, image-to-video, reference-to-video 및 `-fast-*` 변형)의 경우, 네 가지 워크플로 모델(Reference / Edit / Extend / Stitch), 표준 프롬프트 패턴, 멀티모달 입력 제한 및 가격 세부 정보는 [Seedance 2.0 가이드](/guides/media/seedance-2-0)를 참조하세요.
+`seedance-2-0-*` 및 `seedance-2-5-*` 모델(text-to-video, image-to-video, reference-to-video, 그리고 Seedance 2.0 `-fast-*` 변형)은 [Seedance 2.0 & 2.5 가이드](/guides/media/seedance-2-0)에서 네 가지 워크플로(Reference / Edit / Extend / Stitch), 패밀리별 멀티모달 한도, faces-off API 동작, 가격 정보를 확인하세요.
 
-### Seedance 동의
+### Seedance faces-off API
 
-Seedance image- 또는 reference-to-video 요청에 사람의 얼굴이 포함된 경우, API는 정책 텍스트와 함께 `409 needs_consent`를 반환하며, `consents.seedance` 증명과 함께 동일한 요청을 다시 제출합니다. 전체 흐름, 동의 객체, 중복 제거 및 철회에 대해서는 [Seedance 얼굴 미디어 동의 가이드](/guides/media/seedance-face-consent)를 참조하세요.
+공개 Seedance 모델은 페이스 에셋 프로비저닝이나 얼굴 동의 증명을 **지원하지 않습니다**. 얼굴이 포함된 입력은 업스트림에서 거부될 수 있습니다. likeness / 페이스 에셋 생성은 Venice 앱 또는 Studio를 사용하세요. 자세한 내용은 [Seedance 가이드](/guides/media/seedance-2-0)를 참고하세요.
 
 ### 비디오 업스케일링
 
diff --git a/ko/guides/media/seedance-2-0.mdx b/ko/guides/media/seedance-2-0.mdx
index 83f1c878..84f8ded5 100644
--- a/ko/guides/media/seedance-2-0.mdx
+++ b/ko/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "Venice의 Seedance 2.0으로 비디오를 생성·편집·확장·스티치하세요 — 하나의 엔드포인트에서 텍스트, 이미지, 레퍼런스-투-비디오 워크플로를 실제 예제와 함께 제공합니다."
-'og:title': "Seedance 2.0 Guide | Venice API Docs"
-'og:description': "Venice API를 통해 Seedance 2.0으로 비디오를 생성, 편집, 확장, 스티치하세요. 워크플로 패턴, 멀티모달 한도, 가격, 실제 예시."
+title: "Seedance 2.0 & 2.5"
+description: "Venice에서 Seedance 2.0 및 2.5로 영상을 생성·편집·확장·스티치 - 텍스트/이미지/레퍼런스 워크플로, faces-off API, 패밀리별 멀티모달 한도."
+'og:title': "Seedance 2.0 & 2.5 가이드 | Venice API Docs"
+'og:description': "Venice API로 Seedance 2.0 및 2.5 영상을 생성·편집·확장·스티치. 워크플로, faces-off, 멀티모달 한도, 가격, 예제."
 ---
 
-Seedance 2.0은 텍스트, 이미지, 레퍼런스 기반 비디오 생성을 위해 세 가지 변형 패밀리로 Venice에 노출된 플래그십 멀티모달 비디오 모델입니다. **reference-to-video** 변형은 특히 강력합니다: 단일 endpoint와 단일 모델 ID가 **네 가지 구분된 워크플로**(Reference, Edit, Extend, Stitch)를 처리합니다 — 워크플로는 **prompt 형태**에서 추론됩니다.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-이 가이드는 변형들, 표준 prompt와 함께 네 가지 워크플로, 멀티모달 input 한도, 가격, 완전한 `curl` 예시를 다룹니다.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## 변형(Variants)
+<Warning>
+**공개 Seedance API에서는 얼굴을 지원하지 않습니다.** Venice는 공개 Seedance 모델에 대해 페이스 에셋 프로비저닝이나 얼굴 동의 증명(`consents.seedance` / `needs_consent`)을 실행하지 않습니다. 얼굴이 포함된 입력은 업스트림에서 콘텐츠 정책 / 제공자 오류로 거부될 수 있습니다. likeness / 페이스 에셋 생성은 Venice 앱 또는 Studio를 사용하세요.
+</Warning>
+
+## Variants
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | 텍스트 prompt만 |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame(과 선택적으로 last-frame) 이미지 그라운딩 |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | 최대 9장의 레퍼런스 이미지 + 3개의 레퍼런스 비디오 + 3개의 레퍼런스 오디오 도너. Reference / Edit / Extend / Stitch 구동 |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | 더 빠르고, 더 낮은 충실도 등급 |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | 더 빠르고, 더 낮은 충실도 등급 |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | 더 빠르고, 더 낮은 충실도 등급. 동일한 워크플로 세트 |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-모든 변형은 비동기입니다. `POST /api/v1/video/queue`로 제출한 다음, 응답 본문이 `video/mp4`가 될 때까지 `POST /api/v1/video/retrieve`로 폴링하세요. 일반적인 큐 흐름은 [Video Generation](/guides/media/video-generation)을 참고하세요.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## "하나의 모델, 네 가지 워크플로" 모델
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-reference-to-video 변형(`seedance-2-0-reference-to-video`와 그의 Fast 형제)은 네 가지 다른 작업을 처리하는 동일한 기반 모델입니다. **모델은 prompt prefix와 input 형태에서 작업을 추론합니다.** `task`나 `workflow` 필드가 없습니다 — prompt 문법이 라우팅입니다.
+## The "one model, four workflows" model
+
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | 업로드한 레퍼런스 파일을 피사체 / 모션 / 스타일 / 오디오의 도너로 사용해 새 비디오 생성 | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | 텍스트 + 최소 1개의 이미지 또는 비디오 레퍼런스(이미지 0-9개, 비디오 0-3개), 선택적으로 최대 3개 오디오 도너 |
-| **Edit** | 나머지를 보존하면서 하나의 input 비디오를 수정 | `Strictly edit <Video 1>, changing its ...` | 1개의 input 비디오 + 텍스트(이미지 선택적 그라운딩) |
-| **Extend** | 한 클립의 앞으로 / 뒤로 확장 | `Extend <Video 1>, generate ...` | 1개의 input 비디오 + 텍스트 |
-| **Stitch** | 자동 생성 전환으로 2-3개 클립 연결 | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3개의 input 비디오 + 텍스트 |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-**Prompt 문법은 표준이며 대소문자 구분합니다**: 꺾쇠 괄호, 대문자 첫 글자, 숫자 앞 단일 공백 — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## 워크플로 패턴
+## Workflow patterns
 
-### Reference 워크플로
+### Reference workflow
 
-업로드된 레퍼런스 파일을 **도너** — 피사체, 장면, 모션, 스타일, 목소리 음색 — 로 사용해 완전히 새로운 비디오를 생성합니다.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**표준 prompt 패턴**:
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,160 +60,158 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**예시**:
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (오디오 도너는 최소 하나의 이미지 또는 비디오 레퍼런스와 짝지어져야 합니다 — 오디오 단독은 거부됩니다)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Edit 워크플로
+### Edit workflow
 
-하나의 input 비디오를 수정합니다. **prompt에 명시되지 않은 것은 모두 보존됩니다.** 완전히 새로운 비디오가 아닌 국소적 변경(피사체 교체, 날씨/색상 변경, 요소 추가/제거)을 원할 때 사용하세요.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**표준 prompt 패턴**:
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**더 세밀한 제어를 위한 서브 패턴**:
+**Sub-patterns for finer control**:
 
 ```
-요소 추가:
+Add Elements:
   At [timestamp / timing] and [spatial location] of <Video 1>, add [description of intended element].
 
-요소 제거:
+Remove Elements:
   Remove [element to be deleted] from <Video 1>, keeping the rest of the video content unchanged.
 
-요소 수정:
+Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**예시**:
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-마지막 예시는 Edit를 이미지 레퍼런스와 결합합니다 — 완전히 적법하며, 모델은 `<Image 1>`을 교체 대상의 시각적 도너로 사용합니다.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Extend 워크플로
+### Extend workflow
 
-한 클립을 시간상 앞으로 또는 뒤로 이어갑니다. **기본적으로 Seedance는 새 콘텐츠만 반환합니다** — 원본 input이 확장과 함께 이어진 형태가 아닙니다. 이는 전환 연속성을 위한 의도된 동작입니다. input 클립을 확장과 함께 보존하고 싶다면 명시적으로 그렇게 말하세요:
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← 시작에 input 보존
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← 끝에 input 보존
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-전환 처리: 모델이 매끄러운 블렌딩을 위해 전환 프레임을 자동으로 추출하며, input 비디오의 원본 세그먼트는 재생성되지 않습니다.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**예시**:
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch 워크플로(Track Completion)
+### Stitch workflow (Track Completion)
 
-2-3개의 input 클립을 AI 생성 전환으로 연결합니다. 결합된 input 총 길이는 15초 이하여야 합니다.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**표준 prompt 패턴**:
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**예시**:
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-모델은 연속성을 위해 접합점에서 연결 세그먼트를 자동으로 트림합니다.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## 범용 prompt 공식
+## Universal prompt formula
 
-네 가지 워크플로 모두에서 권장되는 작성 공식:
+Across all four workflows, the recommended authoring formula is:
 
 ```
-Subject + Motion + Environment (선택)
-       + Camera Movement / Cut (선택)
-       + Aesthetic Description (선택)
-       + Audio (선택)
+Subject + Motion + Environment (Optional)
+       + Camera Movement / Cut (Optional)
+       + Aesthetic Description (Optional)
+       + Audio (Optional)
 ```
 
-- **Subject + Motion**: 논리적 기반 — "누가" "어떤 행동"을 하는지 정의
-- **Environment + Aesthetics**: 공간 배경, 조명, 시각 스타일
-- **Camera**: 명시적 샷 유형 또는 움직임
-- **Audio**: 몰입감 있는 출력을 위한 환경음 또는 보이스 디렉션
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-이를 워크플로 prefix(예: `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) 위에 레이어링하면 최고 품질의 출력을 만들어냅니다.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## 멀티모달 input 한도
-
-아래 값은 Venice API가 허용하는 값입니다. 이 범위를 벗어나는 요청은 추론에 도달하기 전 스키마 레이어에서 400으로 거부됩니다.
+## Multimodal input limits
 
-### 이미지
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-| Constraint | Value |
-|---|---|
-| Input 방식 | URL(`http://`, `https://`) 또는 Base64 data URL(`data:image/...`) |
-| 포맷 | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| 종횡비(W / H) | 배타적 `(0.4, 2.5)` |
-| 최소 변 | ≥ 300 px |
-| 이미지 수: I2V first-frame | 1 |
-| 이미지 수: I2V first + last frame | 2 |
-| **이미지 수: R2V (V2 / Fast)** | **1 – 9** |
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-### 비디오
+### Shared media floors
 
 | Constraint | Value |
 |---|---|
-| Input 방식 | URL(`http://`, `https://`) 또는 Base64 data URL(`data:video/...`) |
-| 포맷 | `.mp4`, `.mov` |
-| 비디오 코덱 | H.264 / AVC, H.265 / HEVC |
-| 오디오 코덱(컨테이너 내) | AAC, MP3 |
-| **클립당 길이** | `[2, 15]` 초(포함) |
-| **최대 클립 수** | 3 (R2V / Stitch / Extend) |
-| **결합 총 길이** | 모든 클립에 걸쳐 ≤ 15초 |
-| **클립당 크기** | ≤ 50 MB |
-
-### 오디오
-
-| Constraint | Value |
-|---|---|
-| Input 방식 | URL(`http://`, `https://`) 또는 Base64 data URL(`data:audio/...`) |
-| 포맷 | `.wav`, `.mp3` |
-| **클립당 길이** | `[2, 15]` 초 |
-| **최대 클립 수** | 3 |
-| **결합 총 길이** | 모든 클립에 걸쳐 ≤ 15초 |
-| **클립당 크기** | ≤ 15 MB |
-
-레퍼런스 오디오는 R2V 변형에서만 지원됩니다. 각 항목은 모델에 `role: "reference_audio"` 콘텐츠 항목으로 전달되며, prompt는 `<Audio 1>`, `<Audio 2>`, `<Audio 3>`으로 주소를 지정합니다 — 모델은 prompt가 그것을 어떻게 프레이밍하느냐에 따라 각 클립을 보이스 음색, 효과음, 배경 음악에 사용합니다. 레거시 단수 `audio_url` 필드는 같은 콘텐츠 형태로 매핑되며, 이제 한 개 요소 `reference_audio_urls`를 전달하는 것과 동등합니다.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls`만 단독으로 레퍼런스 input일 수는 없습니다.** 모델은 모든 오디오 도너와 함께 최소 하나의 이미지 또는 비디오 레퍼런스가 있어야 합니다. `reference_audio_urls`를 `reference_image_urls`, `reference_video_urls`, `image_url`, `video_url`과 짝지으세요 — 오디오 전용 제출은 거부됩니다.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### 요청 크기
+### Request size
 
-queue endpoint는 최대 **35 MB**의 JSON 본문을 받습니다. 대용량 비디오의 인라인 data URL은 이를 초과할 수 있습니다 — 특히 멀티 클립 Stitch에서는 인라인 base64보다 URL을 선호하세요.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## 가격
+## Pricing
 
-`/video/queue`에 제출하기 전에 주어진 요청 형태에 대한 견적을 받으려면 `POST /api/v1/video/quote`를 호출하세요. quote endpoint가 유일한 권위 있는 출처입니다. 가격 세부 사항은 변경될 수 있으며 클라이언트 측에서 캐시하거나 복제해서는 안 됩니다.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-레퍼런스 비디오가 요청에 포함되어 있으면, 견적이 `/video/queue`가 부과하는 금액과 일치하도록 `reference_video_total_duration`(모든 레퍼런스 클립 길이의 합, 초 단위)도 전달하세요:
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,9 +228,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## 완전한 예시
+## Complete examples
 
-모든 예시는 `VENICE_API_KEY`가 환경에 설정되어 있다고 가정합니다.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
 ### Text-to-video
 
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Image-to-video(first frame)
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video`(와 그의 Fast 변형)는 **`aspect_ratio`를 받지 않습니다** — 출력 종횡비는 입력 이미지의 크기에서 자동 파생됩니다. 필드를 전달하면 *"This model does not support aspect_ratio"* 와 함께 400을 반환합니다. 명시적 종횡비 제어가 필요하면 T2V 또는 R2V 변형을 사용하세요.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Reference 워크플로 — 피사체 도너
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Reference 워크플로 — 피사체 + 오디오 도너
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit 워크플로
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### 이미지 그라운딩이 있는 Edit 워크플로
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -344,7 +385,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch(3개 클립)
+### Stitch (3 clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### 완료 폴링
+### Polling for completion
 
-큐 제출 후 반환된 `queue_id`를 저장하고 응답 본문이 `video/mp4`가 될 때까지 `/video/retrieve`로 폴링하세요:
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-응답은 작업이 완료될 때까지 JSON(`{ "status": "queued" | "running" | "failed", ... }`)이며, 완료되면 응답 본문이 `video/mp4` 바이트로 전환됩니다. 전체 폴링 패턴은 [Video Generation](/guides/media/video-generation)을 참고하세요.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## 문제 해결
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-Reference-to-video 제출에는 `reference_image_urls`, `reference_video_urls`, `image_references`, `video_references` 중 최소 하나가 포함되어야 합니다. 순수 텍스트만의 생성은 유효한 R2V 워크플로가 아닙니다 — 대신 `seedance-2-0-text-to-video`를 사용하세요. `reference_audio_urls`만으로는 충분하지 않습니다(위 Audio 섹션 참고).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-모델은 레퍼런스 비디오를 3개로 제한합니다. 더 많은 클립이 필요하면 먼저 Stitch를 실행(3 → 1)한 다음 그 출력을 후속 작업의 레퍼런스로 사용하세요.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / 합계 `> 15s`
+### Prompt routes to the wrong workflow
 
-클립당 길이는 `[2, 15]` 초 **포함**이며, 모든 레퍼런스 비디오의 합도 15초로 제한됩니다. 제출 전 클라이언트 측에서 클립을 트림하세요.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### Prompt가 잘못된 워크플로로 라우팅됨
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-워크플로는 prompt 문법에서 추론됩니다. 흔한 잘못된 라우팅:
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- **Extend**를 원하지만 `Refer to ...`로 작성 → 모델이 비디오를 이어갈 *캔버스*가 아닌 *도너*로 취급
-- **Stitch**를 원하지만 `Refer to ...`로 작성 → 모델이 하나를 도너로 선택하고 나머지를 무시
-- **Edit**를 원하지만 `Generate a video based on <Video 1>`로 작성 → 모호함, 모델이 Reference로 기본 처리할 수 있음
+### Face / likeness inputs rejected
 
-표준 prefix를 정확히 작성된 대로 사용하세요: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### 견적이 큐 금액과 일치하지 않음
+### Quote doesn't match the queued amount
 
-레퍼런스 비디오를 포함했지만 `reference_video_total_duration`을 `/video/quote`에 전달하지 않았다면 견적과 큐 금액이 다를 수 있습니다. 레퍼런스 비디오가 있을 때는 항상 `reference_video_total_duration`(모든 레퍼런스 클립 길이의 합, 초 단위)을 전달하세요.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## 참고 자료
+## References
 
-- Venice 비디오 큐 endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice 견적 endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- 동반 가이드: [Reference to Video](/guides/media/reference-to-video)(Kling O3 + Grok Imagine R2V 다룸)
-- 동반 가이드: [Video Generation](/guides/media/video-generation)(큐 / 폴링 개요)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/pt-BR/api-reference/endpoint/video/queue.mdx b/pt-BR/api-reference/endpoint/video/queue.mdx
index 9872764a..c261b092 100644
--- a/pt-BR/api-reference/endpoint/video/queue.mdx
+++ b/pt-BR/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Chame `/video/quote` para obter uma estimativa de preço e, em seguida, consulte
 
 Os modelos privados também retornam uma `download_url` para o vídeo finalizado. É uma URL de entrega de curta duração (algumas tentativas são aceitáveis se um download falhar); veja o [guia de geração de vídeo](/guides/media/video-generation#private-download-links) para detalhes e o `DELETE` opcional para privacidade.
 
-### Seedance 2.0
+### Seedance 2.0 e 2.5
 
-Para modelos `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*`), consulte o [Guia do Seedance 2.0](/guides/media/seedance-2-0) para o modelo de quatro workflows (Reference / Edit / Extend / Stitch), padrões canônicos de prompt, limites de entrada multimodal e detalhes de preço.
+Para os modelos `seedance-2-0-*` e `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*` do Seedance 2.0), consulte o [Guia Seedance 2.0 e 2.5](/guides/media/seedance-2-0) para o modelo de quatro fluxos (Reference / Edit / Extend / Stitch), limites multimodais por família, comportamento faces-off da API e detalhes de preços.
 
-### Consentimento Seedance
+### API Seedance faces-off
 
-Quando uma requisição image- ou reference-to-video do Seedance contém um rosto humano, a API retorna `409 needs_consent` com o texto da política, e você reenvia a mesma requisição com uma atestação `consents.seedance`. Veja o [Guia de consentimento para mídia facial do Seedance](/guides/media/seedance-face-consent) para o fluxo completo, o objeto de consentimento, deduplicação e revogação.
+Os modelos Seedance públicos **não** suportam provisionamento de face assets nem atestação de consentimento facial. Entradas com rostos podem ser rejeitadas upstream. Para geração com likeness / face assets, use o app ou o Studio da Venice. Veja o [guia Seedance](/guides/media/seedance-2-0).
 
 ### Upscaling de vídeo
 
diff --git a/pt-BR/guides/media/seedance-2-0.mdx b/pt-BR/guides/media/seedance-2-0.mdx
index e96bfa21..2e7117ec 100644
--- a/pt-BR/guides/media/seedance-2-0.mdx
+++ b/pt-BR/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "Gere, edite, estenda e costure vídeos com o Seedance 2.0 na Venice — fluxos de texto, imagem e reference-to-video em um único endpoint, com exemplos práticos."
-'og:title': "Guia do Seedance 2.0 | Documentação da API Venice"
-'og:description': "Gere, edite, estenda e costure vídeos com o Seedance 2.0 via a API Venice. Padrões de workflow, limites multimodais, preços e exemplos completos."
+title: "Seedance 2.0 e 2.5"
+description: "Gere, edite, estenda e una vídeos com Seedance 2.0 e 2.5 na Venice - fluxos texto/imagem/referência, API faces-off e limites multimodais por família."
+'og:title': "Guia Seedance 2.0 e 2.5 | Venice API Docs"
+'og:description': "Gere, edite, estenda e una vídeos com Seedance 2.0 e 2.5 via API Venice. Fluxos, faces-off, limites, preços e exemplos."
 ---
 
-O Seedance 2.0 é um modelo multimodal de vídeo flagship exposto na Venice como uma família de três variantes para geração de vídeo guiada por texto, imagem e referência. A variante **reference-to-video** é excepcionalmente poderosa: um único endpoint e um único ID de modelo cuidam de **quatro workflows distintos** (Reference, Edit, Extend, Stitch) — o workflow é inferido pelo **formato do seu prompt**.
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-Este guia percorre as variantes, os quatro workflows com seus prompts canônicos, os limites de entrada multimodais, preços e exemplos completos em `curl`.
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## Variantes
+<Warning>
+**Rostos não são suportados na API pública do Seedance.** A Venice não executa provisionamento de face assets nem atestação de consentimento facial (`consents.seedance` / `needs_consent`) para modelos Seedance públicos. Entradas com rostos podem ser rejeitadas upstream como erro de política de conteúdo / provedor. Para geração com likeness / face assets, use o app ou o Studio da Venice.
+</Warning>
+
+## Variants
 
-| ID do modelo | Variante | Resoluções de saída | Notas |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Apenas prompt de texto |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Grounding por imagem do primeiro frame (e opcionalmente último frame) |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Até 9 imagens de referência + 3 vídeos de referência + 3 áudios de referência. Alimenta Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Tier mais rápido e de menor fidelidade |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Tier mais rápido e de menor fidelidade |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Tier mais rápido e de menor fidelidade; mesmo conjunto de workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-Todas as variantes são assíncronas. Envie via `POST /api/v1/video/queue`, e depois faça polling em `POST /api/v1/video/retrieve` até o corpo da resposta ser `video/mp4`. Veja [Geração de vídeo](/guides/media/video-generation) para o fluxo geral de fila.
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## O modelo "um modelo, quatro workflows"
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-A variante reference-to-video (`seedance-2-0-reference-to-video` e sua versão Fast) é o mesmo modelo subjacente servindo quatro tarefas diferentes. **O modelo infere a tarefa a partir do prefixo do prompt e do formato das suas entradas.** Não há campo `task` ou `workflow` — a sintaxe do prompt é o roteamento.
+## The "one model, four workflows" model
 
-| Workflow | O que faz | Prefixo do prompt | Entradas |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Gera um novo vídeo usando arquivos de referência enviados como doadores para sujeito / movimento / estilo / áudio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Texto + ≥1 imagem OU vídeo de referência (0-9 imagens, 0-3 vídeos), mais opcionalmente até 3 doadores de áudio |
-| **Edit** | Modifica um único vídeo de entrada preservando o resto | `Strictly edit <Video 1>, changing its ...` | 1 vídeo de entrada + texto (imagens são grounding opcional) |
-| **Extend** | Extensão para frente / para trás de um clipe | `Extend <Video 1>, generate ...` | 1 vídeo de entrada + texto |
-| **Stitch** | Costura 2-3 clipes com transições geradas automaticamente | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 vídeos de entrada + texto |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-A **sintaxe do prompt é canônica e sensível a maiúsculas**: colchetes angulares, primeira letra maiúscula, um único espaço antes do número — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Padrões de workflow
+## Workflow patterns
 
-### Workflow Reference
+### Reference workflow
 
-Use os arquivos de referência enviados como **doadores** — sujeito, cena, movimento, estilo, timbre vocal — para gerar um vídeo totalmente novo.
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**Padrões de prompt canônicos**:
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,23 +60,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Exemplos**:
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (doadores de áudio devem ser pareados com pelo menos uma referência de imagem ou vídeo — apenas áudio é rejeitado)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Workflow Edit
+### Edit workflow
 
-Modifique um único vídeo de entrada. **Qualquer coisa que não seja explicitamente nomeada no prompt é preservada.** Use isso quando você quer uma mudança localizada (troca de sujeito, mudança de clima/cor, adição/remoção de elemento) em vez de um vídeo totalmente novo.
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**Padrão de prompt canônico**:
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Subpadrões para controle mais fino**:
+**Sub-patterns for finer control**:
 
 ```
 Add Elements:
@@ -80,57 +89,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Exemplos**:
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-O último exemplo combina Edit com uma referência de imagem — perfeitamente legal, o modelo usa `<Image 1>` como doador visual para a substituição.
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Workflow Extend
+### Extend workflow
 
-Continue um único clipe para frente ou para trás no tempo. **Por padrão, o Seedance retorna apenas o novo conteúdo** — não o input original concatenado com a extensão. Isso é por design, para continuidade da transição; se você quiser que o clipe de input seja preservado junto com a extensão, diga isso explicitamente:
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserva input no início
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserva input no fim
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-Tratamento de transição: o modelo extrai automaticamente os frames de transição para mesclagem suave, e os segmentos originais do vídeo de input não são regerados.
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**Exemplos**:
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Workflow Stitch (Track Completion)
+### Stitch workflow (Track Completion)
 
-Conecte 2-3 clipes de input com transições geradas por IA. A duração total combinada de input deve ser ≤ 15 s.
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**Padrão de prompt canônico**:
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Exemplos**:
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-O modelo apara automaticamente os segmentos de conexão nos pontos de junção para continuidade.
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## Fórmula de prompt universal
+## Universal prompt formula
 
-Em todos os quatro workflows, a fórmula de autoria recomendada é:
+Across all four workflows, the recommended authoring formula is:
 
 ```
 Subject + Motion + Environment (Optional)
@@ -139,72 +148,70 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**: a base lógica — define "Quem" está realizando "Qual ação"
-- **Environment + Aesthetics**: plano de fundo espacial, iluminação, estilo visual
-- **Camera**: tipo de plano ou movimento explícito
-- **Audio**: efeitos sonoros ambientes ou direção vocal para saída imersiva
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-Sobrepor isso ao prefixo de um workflow (por exemplo, `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produz as saídas de mais alta qualidade.
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## Limites de entrada multimodal
+## Multimodal input limits
 
-Os valores abaixo são o que a API Venice aceita. Requisições fora dessas faixas são rejeitadas na camada de schema com um 400 antes de chegar à inferência.
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### Imagens
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| Restrição | Valor |
-|---|---|
-| Métodos de input | URL (`http://`, `https://`) ou Base64 data URL (`data:image/...`) |
-| Formatos | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Aspect ratio (W / H) | exclusivo `(0.4, 2.5)` |
-| Lado mínimo | ≥ 300 px |
-| Quantidade de imagem: primeiro frame I2V | 1 |
-| Quantidade de imagem: I2V primeiro + último frame | 2 |
-| **Quantidade de imagem: R2V (V2 / Fast)** | **1 – 9** |
+### Shared media floors
 
-### Vídeos
-
-| Restrição | Valor |
-|---|---|
-| Métodos de input | URL (`http://`, `https://`) ou Base64 data URL (`data:video/...`) |
-| Formatos | `.mp4`, `.mov` |
-| Codecs de vídeo | H.264 / AVC, H.265 / HEVC |
-| Codecs de áudio (no container) | AAC, MP3 |
-| **Duração por clipe** | `[2, 15]` s (inclusivo) |
-| **Máximo de clipes** | 3 (R2V / Stitch / Extend) |
-| **Duração total combinada** | ≤ 15 s em todos os clipes |
-| **Tamanho por clipe** | ≤ 50 MB |
-
-### Áudio
-
-| Restrição | Valor |
+| Constraint | Value |
 |---|---|
-| Métodos de input | URL (`http://`, `https://`) ou Base64 data URL (`data:audio/...`) |
-| Formatos | `.wav`, `.mp3` |
-| **Duração por clipe** | `[2, 15]` s |
-| **Máximo de clipes** | 3 |
-| **Duração total combinada** | ≤ 15 s em todos os clipes |
-| **Tamanho por clipe** | ≤ 15 MB |
-
-Áudio de referência é suportado apenas nas variantes R2V. Cada entrada é encaminhada ao modelo como um item de conteúdo `role: "reference_audio"` que o prompt endereça como `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — o modelo usa cada clipe para timbre vocal, efeitos sonoros ou música de fundo dependendo de como o prompt o enquadra. O campo legado singular `audio_url` mapeia para o mesmo formato de conteúdo e agora é equivalente a passar um `reference_audio_urls` com um único elemento.
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` não pode ser a única entrada de referência.** O modelo requer pelo menos uma referência de imagem ou vídeo junto com qualquer doador de áudio. Pareie `reference_audio_urls` com `reference_image_urls`, `reference_video_urls`, `image_url` ou `video_url` — envios apenas de áudio são rejeitados.
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### Tamanho da requisição
+### Request size
 
-O endpoint de queue aceita corpos JSON de até **35 MB**. Data URLs inline para vídeos grandes podem ultrapassar isso — para Stitch multi-clipe em particular, prefira URLs em vez de base64 inline.
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## Preços
+## Pricing
 
-Chame `POST /api/v1/video/quote` para obter uma cotação para um determinado formato de requisição antes de enviá-la a `/video/queue`. O endpoint de cotação é a única fonte oficial; detalhes de preços podem mudar e não devem ser cacheados ou duplicados no cliente.
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-Quando vídeo(s) de referência fizerem parte da requisição, também passe `reference_video_total_duration` (a soma de todas as durações de clipes de referência em segundos) para que a cotação corresponda ao que `/video/queue` vai cobrar:
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,9 +228,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Exemplos completos
+## Complete examples
 
-Todos os exemplos assumem que `VENICE_API_KEY` está definido no ambiente.
+All examples assume `VENICE_API_KEY` is set in the environment.
 
 ### Text-to-video
 
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Image-to-video (primeiro frame)
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video` (e sua variante Fast) **não aceitam `aspect_ratio`** — o aspect ratio de saída é auto-derivado das dimensões da imagem de input. Passar o campo retorna um 400 com *"This model does not support aspect_ratio"*. Use as variantes T2V ou R2V se precisar de controle explícito de aspect ratio.
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Workflow Reference — doador de sujeito
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Reference — doador de sujeito + áudio
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Edit
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Workflow Edit com grounding de imagem
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend para frente
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -344,7 +385,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clipes)
+### Stitch (3 clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling para conclusão
+### Polling for completion
 
-Após cada envio à fila, salve o `queue_id` retornado e faça polling em `/video/retrieve` até o corpo da resposta ser `video/mp4`:
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-A resposta é JSON (`{ "status": "queued" | "running" | "failed", ... }`) até o job ser concluído, momento em que o corpo da resposta passa para bytes `video/mp4`. Veja [Geração de vídeo](/guides/media/video-generation) para o padrão completo de polling.
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## Solução de problemas
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-Envios de reference-to-video devem incluir pelo menos um de `reference_image_urls`, `reference_video_urls`, `image_references` ou `video_references`. Geração apenas com texto não é um workflow R2V válido — use `seedance-2-0-text-to-video` em vez disso. `reference_audio_urls` sozinho não é suficiente (veja a seção Áudio acima).
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-O modelo limita os vídeos de referência a 3. Se precisar de mais clipes, execute um Stitch primeiro (3 → 1) e depois use a saída como referência para uma sequência.
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / agregado `> 15s`
+### Prompt routes to the wrong workflow
 
-Duração por clipe é `[2, 15]` segundos **inclusivo**; a soma em todos os vídeos de referência também é limitada a 15 segundos. Apare os clipes no cliente antes do envio.
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### O prompt vai para o workflow errado
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-O workflow é inferido pela sintaxe do prompt. Erros comuns de roteamento:
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- Querer **Extend** mas escrever `Refer to ...` → o modelo trata seu vídeo como um *doador*, não como um canvas a continuar
-- Querer **Stitch** mas escrever `Refer to ...` → o modelo escolhe um como doador, ignora os outros
-- Querer **Edit** mas escrever `Generate a video based on <Video 1>` → ambíguo; o modelo pode optar por Reference
+### Face / likeness inputs rejected
 
-Use os prefixos canônicos exatamente como escritos: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### A cotação não corresponde ao valor enfileirado
+### Quote doesn't match the queued amount
 
-Se você incluiu um vídeo de referência mas não passou `reference_video_total_duration` para `/video/quote`, a cotação e o valor enfileirado podem diferir. Sempre passe `reference_video_total_duration` (soma de todas as durações dos clipes de referência, em segundos) quando vídeos de referência estiverem presentes.
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## Referências
+## References
 
-- Endpoint de fila de vídeo da Venice: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Endpoint de cotação da Venice: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Guia companheiro: [Reference to Video](/guides/media/reference-to-video) (cobre Kling O3 + Grok Imagine R2V)
-- Guia companheiro: [Geração de vídeo](/guides/media/video-generation) (visão geral de fila / polling)
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
diff --git a/zh/api-reference/endpoint/video/queue.mdx b/zh/api-reference/endpoint/video/queue.mdx
index e1852398..a1fe02a7 100644
--- a/zh/api-reference/endpoint/video/queue.mdx
+++ b/zh/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ openapi: 'POST /video/queue'
 
 私有模型还会为已完成的视频返回 `download_url`。这是一个短期分发 URL（下载失败时几次重试是可以的）；详情和可选的 `DELETE` 隐私操作请参阅[视频生成指南](/guides/media/video-generation#private-download-links)。
 
-### Seedance 2.0
+### Seedance 2.0 与 2.5
 
-对于 `seedance-2-0-*` 模型（文生视频、图生视频、参考生视频以及 `-fast-*` 变体），请参阅 [Seedance 2.0 指南](/guides/media/seedance-2-0) 了解四工作流模型（Reference / Edit / Extend / Stitch）、规范的 prompt 模式、多模态输入限制和定价详情。
+对于 `seedance-2-0-*` 与 `seedance-2-5-*` 模型（text-to-video、image-to-video、reference-to-video，以及 Seedance 2.0 的 `-fast-*` 变体），请参阅 [Seedance 2.0 与 2.5 指南](/guides/media/seedance-2-0)，了解四工作流模型（Reference / Edit / Extend / Stitch）、按系列区分的多模态限制、faces-off API 行为与定价说明。
 
-### Seedance 同意声明
+### Seedance faces-off API
 
-当 Seedance 的图生视频或参考生视频请求包含人脸时，API 会返回带有政策文本的 `409 needs_consent`，您需要在重新提交相同请求时附带 `consents.seedance` 证明。完整流程、同意对象、去重和撤销机制请参阅 [Seedance 人脸媒体同意指南](/guides/media/seedance-face-consent)。
+公共 Seedance 模型**不**支持人脸资产开通或人脸同意证明。含人脸的输入可能会在上游被拒绝。如需 likeness / 人脸资产生成，请使用 Venice 应用或 Studio。详见 [Seedance 指南](/guides/media/seedance-2-0)。
 
 ### 视频放大
 
diff --git a/zh/guides/media/seedance-2-0.mdx b/zh/guides/media/seedance-2-0.mdx
index 88dee047..d596232c 100644
--- a/zh/guides/media/seedance-2-0.mdx
+++ b/zh/guides/media/seedance-2-0.mdx
@@ -1,49 +1,58 @@
 ---
-title: "Seedance 2.0"
-description: "在 Venice 上使用 Seedance 2.0 生成、编辑、延展并拼接视频 —— 在同一个端点上提供文本、图像和 reference-to-video 工作流，并附带实战示例。"
-'og:title': "Seedance 2.0 Guide | Venice API Docs"
-'og:description': "Generate, edit, extend and stitch videos with Seedance 2.0 via the Venice API. Workflow patterns, multimodal limits, pricing, and worked examples."
+title: "Seedance 2.0 与 2.5"
+description: "使用 Venice 上的 Seedance 2.0 与 2.5 生成、编辑、延长并拼接视频 - 文本/图像/参考工作流、faces-off API 行为，以及按系列区分的多模态限制。"
+'og:title': "Seedance 2.0 与 2.5 指南 | Venice API Docs"
+'og:description': "通过 Venice API 使用 Seedance 2.0 与 2.5 生成、编辑、延长并拼接视频。工作流、faces-off、多模态限制、定价与示例。"
 ---
 
-Seedance 2.0 是一个旗舰多模态视频模型，Venice 将其暴露为三个变体的系列，用于文本驱动、图像驱动和参考驱动的视频生成。**reference-to-video** 变体异常强大：单个端点和单个模型 ID 处理**四种不同的工作流**（Reference、Edit、Extend、Stitch）——工作流从您的 **prompt 形态**推断而出。
+Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
 
-本指南介绍各变体、四种工作流及其规范 prompt、多模态输入限制、定价和完整的 `curl` 示例。
+This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
 
-## 变体
+<Warning>
+**公共 Seedance API 不支持人脸。** Venice 不会为公共 Seedance 模型运行人脸资产开通或人脸同意证明（`consents.seedance` / `needs_consent`）。含人脸的输入可能在上游被以内容政策 / 提供商错误拒绝。如需 likeness / 人脸资产生成，请使用 Venice 应用或 Studio。
+</Warning>
+
+## Variants
 
-| 模型 ID | 变体 | 输出分辨率 | 备注 |
+| Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | 仅文本 prompt |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | 第一帧（及可选最后一帧）图像基础 |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | 最多 9 张参考图像 + 3 段参考视频 + 3 段参考音频供体。驱动 Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | 更快、较低保真度层级 |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | 更快、较低保真度层级 |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | 更快、较低保真度层级；相同工作流集 |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
-所有变体都是异步的。通过 `POST /api/v1/video/queue` 提交，然后轮询 `POST /api/v1/video/retrieve` 直到响应体为 `video/mp4`。常规队列流程请参阅[视频生成](/guides/media/video-generation)。
+All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-## "一个模型，四个工作流"模型
+Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
 
-reference-to-video 变体（`seedance-2-0-reference-to-video` 及其 Fast 兄弟）是同一个底层模型服务四种不同任务。**模型从 prompt 前缀和您输入的形态推断任务。** 没有 `task` 或 `workflow` 字段——prompt 语法即路由。
+## The "one model, four workflows" model
 
-| 工作流 | 功能 | Prompt 前缀 | 输入 |
+The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+
+| Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | 使用上传的参考文件作为主体/动作/风格/音频的供体生成新视频 | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | 文本 + ≥1 张图像 或 视频参考（0-9 张图像，0-3 段视频），可选最多 3 段音频供体 |
-| **Edit** | 修改单个输入视频同时保留其余部分 | `Strictly edit <Video 1>, changing its ...` | 1 段输入视频 + 文本（图像可选基础） |
-| **Extend** | 单个剪辑的前向/后向扩展 | `Extend <Video 1>, generate ...` | 1 段输入视频 + 文本 |
-| **Stitch** | 用自动生成的过渡拼接 2-3 段剪辑 | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 段输入视频 + 文本 |
+| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
+| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
+| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
+| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
 
-**prompt 语法是规范的且区分大小写**：尖括号、首字母大写、数字前单个空格——`<Video 1>`、`<Image 1>`、`<Audio 1>`。
+The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## 工作流模式
+## Workflow patterns
 
-### Reference 工作流
+### Reference workflow
 
-将上传的参考文件用作**供体**——主体、场景、动作、风格、人声音色——以生成全新的视频。
+Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
 
-**规范 prompt 模式**：
+**Canonical prompt patterns**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -51,23 +60,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**示例**：
+**Examples**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.`（音频供体必须与至少一张图像或视频参考配对——仅音频会被拒绝）
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
 
-### Edit 工作流
+### Edit workflow
 
-修改单个输入视频。**prompt 中未明确命名的任何内容都会被保留。** 当您想要局部更改（主体替换、天气/颜色变化、元素添加/移除）而非全新视频时使用此项。
+Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
 
-**规范 prompt 模式**：
+**Canonical prompt pattern**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**更精细控制的子模式**：
+**Sub-patterns for finer control**:
 
 ```
 Add Elements:
@@ -80,57 +89,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**示例**：
+**Examples**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-最后一个示例结合了 Edit 和图像参考——完全合法，模型使用 `<Image 1>` 作为替换的视觉供体。
+The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
 
-### Extend 工作流
+### Extend workflow
 
-将单个剪辑在时间上向前或向后延续。**默认情况下 Seedance 仅返回新内容**——而不是与扩展连接的原始输入。这是出于设计目的，用于过渡连续性；如果您希望保留输入剪辑与扩展一起，请明确说明：
+Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← 在开头保留输入
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← 在结尾保留输入
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
 ```
 
-过渡处理：模型自动提取过渡帧进行无缝混合，输入视频的原始片段不会被重新生成。
+Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
 
-**示例**：
+**Examples**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch 工作流（Track Completion）
+### Stitch workflow (Track Completion)
 
-用 AI 生成的过渡连接 2-3 段输入剪辑。总组合输入时长必须 ≤ 15 s。
+Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
 
-**规范 prompt 模式**：
+**Canonical prompt pattern**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**示例**：
+**Examples**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-模型在连接点自动修剪连接片段以保持连续性。
+The model auto-trims connecting segments at the join points for continuity.
 
 ---
 
-## 通用 prompt 公式
+## Universal prompt formula
 
-在所有四种工作流中，推荐的撰写公式为：
+Across all four workflows, the recommended authoring formula is:
 
 ```
 Subject + Motion + Environment (Optional)
@@ -139,72 +148,70 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**：逻辑基础——定义"谁"在执行"什么动作"
-- **Environment + Aesthetics**：空间背景、光照、视觉风格
-- **Camera**：明确的镜头类型或运动
-- **Audio**：用于沉浸式输出的环境音效或人声方向
+- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
+- **Environment + Aesthetics**: spatial background, lighting, visual style
+- **Camera**: explicit shot type or movement
+- **Audio**: ambient sound effects or vocal direction for immersive output
 
-将其叠加在工作流前缀之上（例如 `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`）能产生最高质量的输出。
+Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
 
 ---
 
-## 多模态输入限制
+## Multimodal input limits
 
-下面的值是 Venice API 接受的内容。超出这些范围的请求在到达推理之前会在 schema 层以 400 拒绝。
+Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
 
-### 图像
+**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
 
-| 约束 | 值 |
-|---|---|
-| 输入方法 | URL（`http://`、`https://`）或 Base64 data URL（`data:image/...`） |
-| 格式 | `.jpeg`、`.png`、`.webp`、`.bmp`、`.tiff`、`.gif`、`.heic`、`.heif` |
-| 宽高比（W / H） | 开区间 `(0.4, 2.5)` |
-| 最小边 | ≥ 300 px |
-| 图像数量：I2V 第一帧 | 1 |
-| 图像数量：I2V 第一 + 最后一帧 | 2 |
-| **图像数量：R2V（V2 / Fast）** | **1 – 9** |
+### Shared media floors
 
-### 视频
-
-| 约束 | 值 |
-|---|---|
-| 输入方法 | URL（`http://`、`https://`）或 Base64 data URL（`data:video/...`） |
-| 格式 | `.mp4`、`.mov` |
-| 视频编解码器 | H.264 / AVC、H.265 / HEVC |
-| 音频编解码器（容器内） | AAC、MP3 |
-| **每个剪辑时长** | `[2, 15]` s（含端点） |
-| **最大剪辑数量** | 3（R2V / Stitch / Extend） |
-| **总组合时长** | 所有剪辑 ≤ 15 s |
-| **每个剪辑大小** | ≤ 50 MB |
-
-### 音频
-
-| 约束 | 值 |
+| Constraint | Value |
 |---|---|
-| 输入方法 | URL（`http://`、`https://`）或 Base64 data URL（`data:audio/...`） |
-| 格式 | `.wav`、`.mp3` |
-| **每个剪辑时长** | `[2, 15]` s |
-| **最大剪辑数量** | 3 |
-| **总组合时长** | 所有剪辑 ≤ 15 s |
-| **每个剪辑大小** | ≤ 15 MB |
-
-参考音频仅在 R2V 变体上支持。每个条目作为 `role: "reference_audio"` 内容项转发到模型，prompt 中以 `<Audio 1>`、`<Audio 2>`、`<Audio 3>` 寻址——模型根据 prompt 框架将每个剪辑用于人声音色、音效或背景音乐。旧版单一 `audio_url` 字段映射到相同的内容形态，现在等同于传递一元素的 `reference_audio_urls`。
+| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
+| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
+| Image minimum side | ≥ 300 px |
+| Video formats | `.mp4`, `.mov` |
+| Video codecs | H.264 / AVC, H.265 / HEVC |
+| Audio codecs (in container) | AAC, MP3 |
+| Audio formats (reference audio) | `.wav`, `.mp3` |
+| Per-clip video size | ≤ 50 MB |
+| Per-clip audio size | ≤ 15 MB |
+| I2V first-frame images | 1 |
+| I2V first + last frame | 2 |
+
+### Family comparison
+
+| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
+|---|---|---|
+| Output duration | 4–15 s | 4–30 s (default 10) |
+| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| R2V reference images | 1–9 | 1–30 |
+| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
+| R2V reference videos | ≤ 3 | ≤ 10 |
+| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
+| Combined reference video duration | ≤ 15 s | ≤ 30 s |
+| R2V reference audio clips | ≤ 3 | ≤ 10 |
+| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
+| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
+
+Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` 不能作为唯一的参考输入。** 模型要求在任何音频供体之外至少有一个图像或视频参考。将 `reference_audio_urls` 与 `reference_image_urls`、`reference_video_urls`、`image_url` 或 `video_url` 配对——纯音频提交会被拒绝。
+**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
 </Warning>
 
-### 请求大小
+### Request size
 
-队列端点接受最大 **35 MB** 的 JSON 正文。大型视频的内联 data URL 可能会超过此值——尤其是多剪辑 Stitch，建议使用 URL 而非内联 base64。
+The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
 
 ---
 
-## 定价
+## Pricing
 
-在提交到 `/video/queue` 之前，调用 `POST /api/v1/video/quote` 获取给定请求形态的报价。报价端点是唯一权威来源；定价细节可能会变化，不应在客户端缓存或复制。
+Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
 
-当请求包含参考视频时，还要传递 `reference_video_total_duration`（所有参考剪辑时长的总秒数），以便报价与 `/video/queue` 实际收费匹配：
+When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
@@ -221,9 +228,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## 完整示例
+## Complete examples
 
-所有示例假设环境中设置了 `VENICE_API_KEY`。
+All examples assume `VENICE_API_KEY` is set in the environment.
 
 ### Text-to-video
 
@@ -240,7 +247,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Image-to-video（第一帧）
+### Seedance 2.5 text-to-video (longer duration)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-text-to-video",
+    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
+    "duration": "20s",
+    "aspect_ratio": "16:9",
+    "resolution": "720p"
+  }'
+```
+
+### Image-to-video (first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -256,10 +278,10 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-`seedance-2-0-image-to-video`（及其 Fast 变体）**不接受 `aspect_ratio`** ——输出宽高比从输入图像的尺寸自动派生。传递该字段会返回 400 错误，并显示 *"This model does not support aspect_ratio"*。如果需要显式宽高比控制，请使用 T2V 或 R2V 变体。
+Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
-### Reference 工作流——主体供体
+### Reference workflow — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -275,7 +297,26 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Reference 工作流——主体 + 音频供体
+### Seedance 2.5 reference workflow — multi-image
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-5-reference-to-video",
+    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
+    "reference_image_urls": [
+      "https://example.com/character.png",
+      "https://example.com/style-board.png"
+    ],
+    "duration": "12s",
+    "aspect_ratio": "9:16",
+    "resolution": "720p"
+  }'
+```
+
+### Reference workflow — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -292,7 +333,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit 工作流
+### Edit workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -309,7 +350,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### 带图像基础的 Edit 工作流
+### Edit workflow with image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -327,7 +368,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### 向前 Extend
+### Extend forward
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -344,7 +385,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch（3 段剪辑）
+### Stitch (3 clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
@@ -365,9 +406,9 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### 轮询完成
+### Polling for completion
 
-每次队列提交后，保存返回的 `queue_id` 并轮询 `/video/retrieve` 直到响应体为 `video/mp4`：
+After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
@@ -380,43 +421,50 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -o output.mp4
 ```
 
-直到作业完成，响应都是 JSON（`{ "status": "queued" | "running" | "failed", ... }`），完成时响应体切换为 `video/mp4` 字节。完整轮询模式请参阅[视频生成](/guides/media/video-generation)。
+The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
 
 ---
 
-## 故障排查
+## Troubleshooting
 
 ### `At least one reference is required for this model`
 
-Reference-to-video 提交必须包含 `reference_image_urls`、`reference_video_urls`、`image_references` 或 `video_references` 中的至少一个。纯文本生成不是有效的 R2V 工作流——请改用 `seedance-2-0-text-to-video`。仅 `reference_audio_urls` 不够（请参阅上面的音频章节）。
+Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
+
+### Too many reference videos / images
+
+Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
+
+### Duration / aggregate duration errors
 
-### `reference_video_urls must have at most 3 videos`
+- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
+- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
 
-模型将参考视频上限设为 3。如果您需要更多剪辑，先运行一次 Stitch（3 → 1），然后将输出用作后续的参考。
+Trim clips client-side before submission.
 
-### `Per clip must be 2–15s` / 聚合 `> 15s`
+### Prompt routes to the wrong workflow
 
-每个剪辑时长是**含端点**的 `[2, 15]` 秒；所有参考视频的总和也上限为 15 秒。在提交之前在客户端修剪剪辑。
+Workflow is inferred from prompt syntax. Common misroutings:
 
-### Prompt 路由到错误的工作流
+- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
+- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
+- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
 
-工作流从 prompt 语法推断。常见的错误路由：
+Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-- 想要 **Extend** 但写 `Refer to ...` → 模型将您的视频视为*供体*，而非要延续的画布
-- 想要 **Stitch** 但写 `Refer to ...` → 模型挑选一个作为供体，忽略其余的
-- 想要 **Edit** 但写 `Generate a video based on <Video 1>` → 含糊；模型可能默认 Reference
+### Face / likeness inputs rejected
 
-完全按所写使用规范前缀：`Strictly edit <Video 1>, ...`、`Extend <Video 1>, ...`、`<Video 1> + ... + followed by <Video 2>`。
+Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
 
-### 报价与队列金额不匹配
+### Quote doesn't match the queued amount
 
-如果您包含了参考视频但没有将 `reference_video_total_duration` 传递给 `/video/quote`，报价和队列金额可能不同。当存在参考视频时，始终传递 `reference_video_total_duration`（所有参考剪辑时长的总秒数）。
+If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
 
 ---
 
-## 参考
+## References
 
-- Venice 视频队列端点：[`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice 报价端点：[`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- 配套指南：[Reference to Video](/guides/media/reference-to-video)（涵盖 Kling O3 + Grok Imagine R2V）
-- 配套指南：[视频生成](/guides/media/video-generation)（队列/轮询概览）
+- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
+- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)

From a3cdc0348de0e2a42e91a46802a923edc8fd32e0 Mon Sep 17 00:00:00 2001
From: Tomas Bystrican <tom.bystrican@gmail.com>
Date: Sat, 8 Aug 2026 18:37:57 +0930
Subject: [PATCH 2/5] Document Seedance 2.0 4K output resolutions (API-103).
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Non-Fast 2.0 variants accept 480p–4k; Fast and 2.5 stay 480p/720p.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 ar/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 de/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 es/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 fr/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 guides/media/seedance-2-0.mdx       | 31 +++++++++++++++++++++--------
 it/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 ko/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 pt-BR/guides/media/seedance-2-0.mdx | 31 +++++++++++++++++++++--------
 zh/guides/media/seedance-2-0.mdx    | 31 +++++++++++++++++++++--------
 9 files changed, 207 insertions(+), 72 deletions(-)

diff --git a/ar/guides/media/seedance-2-0.mdx b/ar/guides/media/seedance-2-0.mdx
index a8281281..bd56b873 100644
--- a/ar/guides/media/seedance-2-0.mdx
+++ b/ar/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/de/guides/media/seedance-2-0.mdx b/de/guides/media/seedance-2-0.mdx
index a808e13b..56056213 100644
--- a/de/guides/media/seedance-2-0.mdx
+++ b/de/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/es/guides/media/seedance-2-0.mdx b/es/guides/media/seedance-2-0.mdx
index 7afa70ef..d00b72f0 100644
--- a/es/guides/media/seedance-2-0.mdx
+++ b/es/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/fr/guides/media/seedance-2-0.mdx b/fr/guides/media/seedance-2-0.mdx
index 91e41d03..6bc7ca88 100644
--- a/fr/guides/media/seedance-2-0.mdx
+++ b/fr/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/guides/media/seedance-2-0.mdx b/guides/media/seedance-2-0.mdx
index d53ac7cb..0b14a32f 100644
--- a/guides/media/seedance-2-0.mdx
+++ b/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/it/guides/media/seedance-2-0.mdx b/it/guides/media/seedance-2-0.mdx
index c9f977db..01e08499 100644
--- a/it/guides/media/seedance-2-0.mdx
+++ b/it/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/ko/guides/media/seedance-2-0.mdx b/ko/guides/media/seedance-2-0.mdx
index 84f8ded5..eceb0aaa 100644
--- a/ko/guides/media/seedance-2-0.mdx
+++ b/ko/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/pt-BR/guides/media/seedance-2-0.mdx b/pt-BR/guides/media/seedance-2-0.mdx
index 2e7117ec..26e4f08d 100644
--- a/pt-BR/guides/media/seedance-2-0.mdx
+++ b/pt-BR/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash
diff --git a/zh/guides/media/seedance-2-0.mdx b/zh/guides/media/seedance-2-0.mdx
index d596232c..a056f013 100644
--- a/zh/guides/media/seedance-2-0.mdx
+++ b/zh/guides/media/seedance-2-0.mdx
@@ -17,19 +17,19 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier; same workflow set |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
 | `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
 | `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
 | `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
-Discover live model IDs with `GET /models?type=video` — do not hardcode availability.
+Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
 
 ## The "one model, four workflows" model
 
@@ -185,7 +185,7 @@ Values below are what the Venice API accepts. Requests outside these ranges are
 | Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
 |---|---|---|
 | Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
+| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
 | R2V reference images | 1–9 | 1–30 |
 | Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
 | R2V reference videos | ≤ 3 | ≤ 10 |
@@ -247,6 +247,21 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
+### Seedance 2.0 text-to-video (4K)
+
+```bash
+curl -X POST https://api.venice.ai/api/v1/video/queue \
+  -H "Authorization: Bearer $VENICE_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "seedance-2-0-text-to-video",
+    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
+    "duration": "5s",
+    "aspect_ratio": "16:9",
+    "resolution": "4k"
+  }'
+```
+
 ### Seedance 2.5 text-to-video (longer duration)
 
 ```bash

From e0a7171cd538cc4f7f47bb57decb59955aa573bc Mon Sep 17 00:00:00 2001
From: Tomas Bystrican <tom.bystrican@gmail.com>
Date: Mon, 10 Aug 2026 09:09:14 +0930
Subject: [PATCH 3/5] Document listed Seedance public ids as *-api twins
 (API-103).

Align public docs with ENG-950 Option 3: faces-off models are listed under
the *-api ids; face-consent page left on the original face model names.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 api-reference/endpoint/video/queue.mdx        |  2 +-
 ar/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 ar/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 ar/guides/media/video-generation.mdx          |  2 +-
 de/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 de/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 de/guides/media/video-generation.mdx          |  2 +-
 es/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 es/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 es/guides/media/video-generation.mdx          |  2 +-
 fr/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 fr/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 fr/guides/media/video-generation.mdx          |  2 +-
 guides/integrations/venice-video-harness.mdx  |  4 +-
 guides/media/seedance-2-0.mdx                 | 48 +++++++++----------
 guides/media/video-generation.mdx             |  2 +-
 it/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 it/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 it/guides/media/video-generation.mdx          |  2 +-
 ko/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 ko/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 ko/guides/media/video-generation.mdx          |  2 +-
 pt-BR/api-reference/endpoint/video/queue.mdx  |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 pt-BR/guides/media/seedance-2-0.mdx           | 48 +++++++++----------
 pt-BR/guides/media/video-generation.mdx       |  2 +-
 zh/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 zh/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 zh/guides/media/video-generation.mdx          |  2 +-
 36 files changed, 252 insertions(+), 252 deletions(-)

diff --git a/api-reference/endpoint/video/queue.mdx b/api-reference/endpoint/video/queue.mdx
index c5c974c2..96fd855a 100644
--- a/api-reference/endpoint/video/queue.mdx
+++ b/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Private models also return a `download_url` for the finished video. It is a shor
 
 ### Seedance 2.0 & 2.5
 
-For `seedance-2-0-*` and `seedance-2-5-*` models (text-to-video, image-to-video, reference-to-video, plus the Seedance 2.0 `-fast-*` variants), see the [Seedance 2.0 & 2.5 Guide](/guides/media/seedance-2-0) for the four-workflow model (Reference / Edit / Extend / Stitch), family-specific multimodal limits, faces-off API behavior, and pricing details.
+For `seedance-2-0-*-api` and `seedance-2-5-*-api` models (text-to-video, image-to-video, reference-to-video, plus the Seedance 2.0 `-fast-*` variants), see the [Seedance 2.0 & 2.5 Guide](/guides/media/seedance-2-0) for the four-workflow model (Reference / Edit / Extend / Stitch), family-specific multimodal limits, faces-off API behavior, and pricing details.
 
 ### Seedance faces-off API
 
diff --git a/ar/api-reference/endpoint/video/queue.mdx b/ar/api-reference/endpoint/video/queue.mdx
index bde1d794..d1625e5b 100644
--- a/ar/api-reference/endpoint/video/queue.mdx
+++ b/ar/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ openapi: 'POST /video/queue'
 
 ### Seedance 2.0 و 2.5
 
-لنماذج `seedance-2-0-*` و `seedance-2-5-*` (text-to-video و image-to-video و reference-to-video، بالإضافة إلى متغيّرات `-fast-*` لـ Seedance 2.0)، راجع [دليل Seedance 2.0 و 2.5](/guides/media/seedance-2-0) لنموذج سير العمل الأربعة (Reference / Edit / Extend / Stitch) وحدود الوسائط المتعددة حسب العائلة وسلوك واجهة faces-off وتفاصيل التسعير.
+لنماذج `seedance-2-0-*-api` و `seedance-2-5-*-api` (text-to-video و image-to-video و reference-to-video، بالإضافة إلى متغيّرات `-fast-*` لـ Seedance 2.0)، راجع [دليل Seedance 2.0 و 2.5](/guides/media/seedance-2-0) لنموذج سير العمل الأربعة (Reference / Edit / Extend / Stitch) وحدود الوسائط المتعددة حسب العائلة وسلوك واجهة faces-off وتفاصيل التسعير.
 
 ### واجهة Seedance بدون وجوه (faces-off)
 
diff --git a/ar/guides/integrations/venice-video-harness.mdx b/ar/guides/integrations/venice-video-harness.mdx
index 44e2a326..429d326b 100644
--- a/ar/guides/integrations/venice-video-harness.mdx
+++ b/ar/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ curl -L -o ~/.cache/whisper.cpp/ggml-base.en.bin \
 
 | الدور | النموذج الافتراضي | متى يُستخدم |
 |---|---|---|
-| لقطات الشخصيات (1-2 شخصية) | `seedance-2-0-reference-to-video` | R2V افتراضي مع `reference_image_urls` مسطّحة، ووسوم `@Image`، حتى 15 ثانية، وصوت ستيريو أصلي |
+| لقطات الشخصيات (1-2 شخصية) | `seedance-2-0-reference-to-video-api` | R2V افتراضي مع `reference_image_urls` مسطّحة، ووسوم `@Image`، حتى 15 ثانية، وصوت ستيريو أصلي |
 | لقطات الشخصيات (3+ شخصيات) | `kling-o3-standard-reference-to-video` | بديل احتياطي تلقائي مع `elements` مهيكلة لهوية متعدّدة الشخصيات |
-| تأسيسي / مزاجي / حركة | `seedance-2-0-image-to-video` | بدون شخصيات؛ جودة سينمائية ملحمية، حتى 15 ثانية |
+| تأسيسي / مزاجي / حركة | `seedance-2-0-image-to-video-api` | بدون شخصيات؛ جودة سينمائية ملحمية، حتى 15 ثانية |
 
 يمكن تجاوز هذه القيم لكل مشروع عبر `series.json → videoDefaults`. لاستهداف عائلة غير Seedance (مثل الحسابات التي تفتقر إلى وصول Seedance)، اضبط `videoDefaults` إلى `kling-o3-standard-reference-to-video` و`veo3.1-fast-image-to-video`.
 
diff --git a/ar/guides/media/seedance-2-0.mdx b/ar/guides/media/seedance-2-0.mdx
index bd56b873..e557ad9b 100644
--- a/ar/guides/media/seedance-2-0.mdx
+++ b/ar/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/ar/guides/media/video-generation.mdx b/ar/guides/media/video-generation.mdx
index a8f8afb0..a8996694 100644
--- a/ar/guides/media/video-generation.mdx
+++ b/ar/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ await fetch(`${BASE_URL}/video/complete`, {
 عرض السعر بالدولار الأمريكي.
 
 <Info>
-**معاملات عرض السعر هي مدخلات تسعير فقط.** الحقول التي ترسلها إلى `/video/quote` — بما في ذلك `aspect_ratio` و`audio` و`reference_video_total_duration` — تُستخدم بحتًا لحساب السعر. **لا** تُمرَّر إلى `/video/queue` ولا تؤثر في الفيديو المولَّد. يُقبل إرسال `aspect_ratio` إلى quote حتى بالنسبة للنماذج (مثل `seedance-2-0-image-to-video`) التي ترفض `aspect_ratio` وقت التوليد، لأن نماذج الصورة إلى فيديو تشتق نسبة العرض إلى الارتفاع للإخراج من صورة الإدخال.
+**معاملات عرض السعر هي مدخلات تسعير فقط.** الحقول التي ترسلها إلى `/video/quote` — بما في ذلك `aspect_ratio` و`audio` و`reference_video_total_duration` — تُستخدم بحتًا لحساب السعر. **لا** تُمرَّر إلى `/video/queue` ولا تؤثر في الفيديو المولَّد. يُقبل إرسال `aspect_ratio` إلى quote حتى بالنسبة للنماذج (مثل `seedance-2-0-image-to-video-api`) التي ترفض `aspect_ratio` وقت التوليد، لأن نماذج الصورة إلى فيديو تشتق نسبة العرض إلى الارتفاع للإخراج من صورة الإدخال.
 </Info>
 
 ### تسعير المرجع إلى فيديو
diff --git a/de/api-reference/endpoint/video/queue.mdx b/de/api-reference/endpoint/video/queue.mdx
index 1f19e992..eb5f84ee 100644
--- a/de/api-reference/endpoint/video/queue.mdx
+++ b/de/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Private Modelle geben außerdem eine `download_url` für das fertige Video zurü
 
 ### Seedance 2.0 & 2.5
 
-Für `seedance-2-0-*`- und `seedance-2-5-*`-Modelle (Text-to-Video, Image-to-Video, Reference-to-Video sowie die Seedance-2.0-`-fast-*`-Varianten) siehe den [Seedance-2.0-&-2.5-Guide](/guides/media/seedance-2-0) zu den vier Workflows (Reference / Edit / Extend / Stitch), familienspezifischen Multimodal-Limits, Faces-off-API-Verhalten und Preisen.
+Für `seedance-2-0-*-api`- und `seedance-2-5-*-api`-Modelle (Text-to-Video, Image-to-Video, Reference-to-Video sowie die Seedance-2.0-`-fast-*`-Varianten) siehe den [Seedance-2.0-&-2.5-Guide](/guides/media/seedance-2-0) zu den vier Workflows (Reference / Edit / Extend / Stitch), familienspezifischen Multimodal-Limits, Faces-off-API-Verhalten und Preisen.
 
 ### Seedance Faces-off API
 
diff --git a/de/guides/integrations/venice-video-harness.mdx b/de/guides/integrations/venice-video-harness.mdx
index 146921f0..337d30b1 100644
--- a/de/guides/integrations/venice-video-harness.mdx
+++ b/de/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Die Standardeinstellungen sind bewusst meinungsstark, weil Konsistenz das Ziel i
 
 | Rolle | Standardmodell | Wann verwendet |
 |---|---|---|
-| Charakter-Shots (1–2 Charaktere) | `seedance-2-0-reference-to-video` | Standard-R2V mit flachen `reference_image_urls`, `@Image`-Tags, bis zu 15 s, nativer Stereo-Sound |
+| Charakter-Shots (1–2 Charaktere) | `seedance-2-0-reference-to-video-api` | Standard-R2V mit flachen `reference_image_urls`, `@Image`-Tags, bis zu 15 s, nativer Stereo-Sound |
 | Charakter-Shots (3+ Charaktere) | `kling-o3-standard-reference-to-video` | Auto-Fallback mit strukturierten `elements` für Multi-Charakter-Identität |
-| Establishing / Stimmung / Action | `seedance-2-0-image-to-video` | Keine Charaktere; epische Kino-Qualität, bis zu 15 s |
+| Establishing / Stimmung / Action | `seedance-2-0-image-to-video-api` | Keine Charaktere; epische Kino-Qualität, bis zu 15 s |
 
 Diese sind pro Projekt über `series.json → videoDefaults` überschreibbar. Für eine Nicht-Seedance-Familie (z. B. Accounts ohne Seedance-Zugriff) `videoDefaults` auf `kling-o3-standard-reference-to-video` und `veo3.1-fast-image-to-video` setzen.
 
diff --git a/de/guides/media/seedance-2-0.mdx b/de/guides/media/seedance-2-0.mdx
index 56056213..8443a526 100644
--- a/de/guides/media/seedance-2-0.mdx
+++ b/de/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/de/guides/media/video-generation.mdx b/de/guides/media/video-generation.mdx
index ec9bdd18..7d94d89b 100644
--- a/de/guides/media/video-generation.mdx
+++ b/de/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Exakte Kosten vor der Generierung. Nur die Pricing-Inputs senden (`model`, `dura
 Das Quote ist in USD.
 
 <Info>
-**Quote-Parameter sind ausschließlich Preis-Inputs.** Felder, die Sie an `/video/quote` senden – darunter `aspect_ratio`, `audio` und `reference_video_total_duration` – dienen ausschließlich der Preisberechnung. Sie werden **nicht** an `/video/queue` weitergeleitet und haben keinen Einfluss auf das generierte Video. `aspect_ratio` an Quote zu senden ist auch bei Modellen zulässig (z. B. `seedance-2-0-image-to-video`), die `aspect_ratio` bei der Generierung ablehnen, da Image-to-Video-Modelle das Ausgabe-Seitenverhältnis aus dem Eingabebild ableiten.
+**Quote-Parameter sind ausschließlich Preis-Inputs.** Felder, die Sie an `/video/quote` senden – darunter `aspect_ratio`, `audio` und `reference_video_total_duration` – dienen ausschließlich der Preisberechnung. Sie werden **nicht** an `/video/queue` weitergeleitet und haben keinen Einfluss auf das generierte Video. `aspect_ratio` an Quote zu senden ist auch bei Modellen zulässig (z. B. `seedance-2-0-image-to-video-api`), die `aspect_ratio` bei der Generierung ablehnen, da Image-to-Video-Modelle das Ausgabe-Seitenverhältnis aus dem Eingabebild ableiten.
 </Info>
 
 ### Preisberechnung bei Reference-to-Video
diff --git a/es/api-reference/endpoint/video/queue.mdx b/es/api-reference/endpoint/video/queue.mdx
index e4448fea..f5035f23 100644
--- a/es/api-reference/endpoint/video/queue.mdx
+++ b/es/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Los modelos privados también devuelven un `download_url` para el vídeo termina
 
 ### Seedance 2.0 y 2.5
 
-Para los modelos `seedance-2-0-*` y `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*` de Seedance 2.0), consulta la [Guía de Seedance 2.0 y 2.5](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), límites multimodales por familia, comportamiento faces-off de la API y detalles de precios.
+Para los modelos `seedance-2-0-*-api` y `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*` de Seedance 2.0), consulta la [Guía de Seedance 2.0 y 2.5](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), límites multimodales por familia, comportamiento faces-off de la API y detalles de precios.
 
 ### Seedance API sin rostros (faces-off)
 
diff --git a/es/guides/integrations/venice-video-harness.mdx b/es/guides/integrations/venice-video-harness.mdx
index f4f03a51..e06c82c2 100644
--- a/es/guides/integrations/venice-video-harness.mdx
+++ b/es/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Los defaults del harness son opinionados porque la consistencia es lo importante
 
 | Rol | Modelo predeterminado | Cuándo se usa |
 |---|---|---|
-| Tomas de personaje (1-2 personajes) | `seedance-2-0-reference-to-video` | R2V por defecto con `reference_image_urls` planos, etiquetas `@Image`, hasta 15s, audio estéreo nativo |
+| Tomas de personaje (1-2 personajes) | `seedance-2-0-reference-to-video-api` | R2V por defecto con `reference_image_urls` planos, etiquetas `@Image`, hasta 15s, audio estéreo nativo |
 | Tomas de personaje (3+ personajes) | `kling-o3-standard-reference-to-video` | Fallback automático con `elements` estructurados para identidad multi-personaje |
-| Establecimiento / atmósfera / acción | `seedance-2-0-image-to-video` | Sin personajes; calidad cinematográfica épica, hasta 15s |
+| Establecimiento / atmósfera / acción | `seedance-2-0-image-to-video-api` | Sin personajes; calidad cinematográfica épica, hasta 15s |
 
 Estos son sobrescribibles por proyecto vía `series.json → videoDefaults`. Para apuntar a una familia que no sea Seedance (p. ej., cuentas sin acceso a Seedance), establece `videoDefaults` en `kling-o3-standard-reference-to-video` y `veo3.1-fast-image-to-video`.
 
diff --git a/es/guides/media/seedance-2-0.mdx b/es/guides/media/seedance-2-0.mdx
index d00b72f0..01daca33 100644
--- a/es/guides/media/seedance-2-0.mdx
+++ b/es/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/es/guides/media/video-generation.mdx b/es/guides/media/video-generation.mdx
index b4c89542..2ec1e380 100644
--- a/es/guides/media/video-generation.mdx
+++ b/es/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtén el coste exacto antes de generar. Envía solo entradas de precio (`model`
 La cotización es en USD.
 
 <Info>
-**Los parámetros de cotización son solo entradas de precio.** Los campos que envías a `/video/quote` —incluyendo `aspect_ratio`, `audio` y `reference_video_total_duration`— se usan únicamente para calcular el precio. **No** se reenvían a `/video/queue` y no afectan al vídeo generado. Enviar `aspect_ratio` a quote se acepta incluso para modelos (como `seedance-2-0-image-to-video`) que rechazan `aspect_ratio` en el momento de la generación, ya que los modelos image-to-video derivan la relación de aspecto de salida a partir de la imagen de entrada.
+**Los parámetros de cotización son solo entradas de precio.** Los campos que envías a `/video/quote` —incluyendo `aspect_ratio`, `audio` y `reference_video_total_duration`— se usan únicamente para calcular el precio. **No** se reenvían a `/video/queue` y no afectan al vídeo generado. Enviar `aspect_ratio` a quote se acepta incluso para modelos (como `seedance-2-0-image-to-video-api`) que rechazan `aspect_ratio` en el momento de la generación, ya que los modelos image-to-video derivan la relación de aspecto de salida a partir de la imagen de entrada.
 </Info>
 
 ### Precios de reference-to-video
diff --git a/fr/api-reference/endpoint/video/queue.mdx b/fr/api-reference/endpoint/video/queue.mdx
index 5c21a917..b6bada21 100644
--- a/fr/api-reference/endpoint/video/queue.mdx
+++ b/fr/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Les modèles privés renvoient également un `download_url` pour la vidéo termi
 
 ### Seedance 2.0 et 2.5
 
-Pour les modèles `seedance-2-0-*` et `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*` de Seedance 2.0), voir le [Guide Seedance 2.0 et 2.5](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les limites multimodales par famille, le comportement faces-off de l’API et les tarifs.
+Pour les modèles `seedance-2-0-*-api` et `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*` de Seedance 2.0), voir le [Guide Seedance 2.0 et 2.5](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les limites multimodales par famille, le comportement faces-off de l’API et les tarifs.
 
 ### API Seedance faces-off
 
diff --git a/fr/guides/integrations/venice-video-harness.mdx b/fr/guides/integrations/venice-video-harness.mdx
index e0848ebc..5399da4c 100644
--- a/fr/guides/integrations/venice-video-harness.mdx
+++ b/fr/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Les défauts du harness sont assumés parce que la cohérence est le but. Le rou
 
 | Rôle | Modèle par défaut | Quand utilisé |
 |---|---|---|
-| Plans de personnage (1-2 personnages) | `seedance-2-0-reference-to-video` | R2V par défaut avec `reference_image_urls` plats, tags `@Image`, jusqu'à 15 s, audio stéréo natif |
+| Plans de personnage (1-2 personnages) | `seedance-2-0-reference-to-video-api` | R2V par défaut avec `reference_image_urls` plats, tags `@Image`, jusqu'à 15 s, audio stéréo natif |
 | Plans de personnage (3+ personnages) | `kling-o3-standard-reference-to-video` | Fallback automatique avec `elements` structurés pour l'identité multi-personnages |
-| Plan d'établissement / d'ambiance / d'action | `seedance-2-0-image-to-video` | Pas de personnages ; qualité cinématique épique, jusqu'à 15 s |
+| Plan d'établissement / d'ambiance / d'action | `seedance-2-0-image-to-video-api` | Pas de personnages ; qualité cinématique épique, jusqu'à 15 s |
 
 Ces valeurs sont remplaçables par projet via `series.json → videoDefaults`. Pour cibler une famille non-Seedance (par ex. des comptes qui n'ont pas accès à Seedance), définissez `videoDefaults` à `kling-o3-standard-reference-to-video` et `veo3.1-fast-image-to-video`.
 
diff --git a/fr/guides/media/seedance-2-0.mdx b/fr/guides/media/seedance-2-0.mdx
index 6bc7ca88..683d51a4 100644
--- a/fr/guides/media/seedance-2-0.mdx
+++ b/fr/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/fr/guides/media/video-generation.mdx b/fr/guides/media/video-generation.mdx
index 6f0aa09f..141cf9ae 100644
--- a/fr/guides/media/video-generation.mdx
+++ b/fr/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtenez le coût exact avant la génération. Envoyez uniquement les entrées de
 Le devis est en USD.
 
 <Info>
-**Les paramètres du devis servent uniquement au calcul du prix.** Les champs que vous envoyez à `/video/quote` — y compris `aspect_ratio`, `audio` et `reference_video_total_duration` — sont utilisés uniquement pour calculer le prix. Ils **ne sont pas** transmis à `/video/queue` et n'influencent pas la vidéo générée. Envoyer `aspect_ratio` au quote est accepté même pour les modèles (tels que `seedance-2-0-image-to-video`) qui rejettent `aspect_ratio` au moment de la génération, puisque les modèles image-to-video dérivent le ratio d'aspect de sortie à partir de l'image d'entrée.
+**Les paramètres du devis servent uniquement au calcul du prix.** Les champs que vous envoyez à `/video/quote` — y compris `aspect_ratio`, `audio` et `reference_video_total_duration` — sont utilisés uniquement pour calculer le prix. Ils **ne sont pas** transmis à `/video/queue` et n'influencent pas la vidéo générée. Envoyer `aspect_ratio` au quote est accepté même pour les modèles (tels que `seedance-2-0-image-to-video-api`) qui rejettent `aspect_ratio` au moment de la génération, puisque les modèles image-to-video dérivent le ratio d'aspect de sortie à partir de l'image d'entrée.
 </Info>
 
 ### Tarification reference-to-video
diff --git a/guides/integrations/venice-video-harness.mdx b/guides/integrations/venice-video-harness.mdx
index 1616218b..15dce8a1 100644
--- a/guides/integrations/venice-video-harness.mdx
+++ b/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ The harness defaults are opinionated because consistency is the point. The curre
 
 | Role | Default model | When used |
 |---|---|---|
-| Character shots (1-2 characters) | `seedance-2-0-reference-to-video` | Default R2V with flat `reference_image_urls`, `@Image` tags, up to 15s, native stereo audio |
+| Character shots (1-2 characters) | `seedance-2-0-reference-to-video-api` | Default R2V with flat `reference_image_urls`, `@Image` tags, up to 15s, native stereo audio |
 | Character shots (3+ characters) | `kling-o3-standard-reference-to-video` | Auto-fallback with structured `elements` for multi-character identity |
-| Establishing / mood / action | `seedance-2-0-image-to-video` | No characters; epic cinematic quality, up to 15s |
+| Establishing / mood / action | `seedance-2-0-image-to-video-api` | No characters; epic cinematic quality, up to 15s |
 
 These are overridable per-project via `series.json → videoDefaults`. To target a non-Seedance family (e.g. accounts that lack Seedance access), set `videoDefaults` to `kling-o3-standard-reference-to-video` and `veo3.1-fast-image-to-video`.
 
diff --git a/guides/media/seedance-2-0.mdx b/guides/media/seedance-2-0.mdx
index 0b14a32f..ef3f1337 100644
--- a/guides/media/seedance-2-0.mdx
+++ b/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/guides/media/video-generation.mdx b/guides/media/video-generation.mdx
index a4ecca8f..72d0084d 100644
--- a/guides/media/video-generation.mdx
+++ b/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Get exact cost before generating. Send only pricing inputs (`model`, `duration`,
 Quote is in USD.
 
 <Info>
-**Quote parameters are pricing inputs only.** Fields you send to `/video/quote` — including `aspect_ratio`, `audio`, and `reference_video_total_duration` — are used purely to compute the price. They are **not** forwarded to `/video/queue` and do not affect the generated video. Sending `aspect_ratio` to quote is accepted even for models (such as `seedance-2-0-image-to-video`) that reject `aspect_ratio` at generation time, since image-to-video models derive the output aspect ratio from the input image.
+**Quote parameters are pricing inputs only.** Fields you send to `/video/quote` — including `aspect_ratio`, `audio`, and `reference_video_total_duration` — are used purely to compute the price. They are **not** forwarded to `/video/queue` and do not affect the generated video. Sending `aspect_ratio` to quote is accepted even for models (such as `seedance-2-0-image-to-video-api`) that reject `aspect_ratio` at generation time, since image-to-video models derive the output aspect ratio from the input image.
 </Info>
 
 ### Reference-to-video pricing
diff --git a/it/api-reference/endpoint/video/queue.mdx b/it/api-reference/endpoint/video/queue.mdx
index 9ae50b97..cb62b18b 100644
--- a/it/api-reference/endpoint/video/queue.mdx
+++ b/it/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ I modelli privati restituiscono anche un `download_url` per il video finito. Si
 
 ### Seedance 2.0 e 2.5
 
-Per i modelli `seedance-2-0-*` e `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*` di Seedance 2.0), vedi la [Guida Seedance 2.0 e 2.5](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i limiti multimodali per famiglia, il comportamento faces-off dell’API e i dettagli sui prezzi.
+Per i modelli `seedance-2-0-*-api` e `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*` di Seedance 2.0), vedi la [Guida Seedance 2.0 e 2.5](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i limiti multimodali per famiglia, il comportamento faces-off dell’API e i dettagli sui prezzi.
 
 ### API Seedance faces-off
 
diff --git a/it/guides/integrations/venice-video-harness.mdx b/it/guides/integrations/venice-video-harness.mdx
index c700692e..9684ee68 100644
--- a/it/guides/integrations/venice-video-harness.mdx
+++ b/it/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ I default dell'harness sono opinionated perché la consistenza è il punto chiav
 
 | Ruolo | Modello di default | Quando usato |
 |---|---|---|
-| Inquadrature con personaggi (1-2 personaggi) | `seedance-2-0-reference-to-video` | R2V di default con `reference_image_urls` flat, tag `@Image`, fino a 15s, audio stereo nativo |
+| Inquadrature con personaggi (1-2 personaggi) | `seedance-2-0-reference-to-video-api` | R2V di default con `reference_image_urls` flat, tag `@Image`, fino a 15s, audio stereo nativo |
 | Inquadrature con personaggi (3+ personaggi) | `kling-o3-standard-reference-to-video` | Fallback automatico con `elements` strutturati per identità multi-personaggio |
-| Establishing / mood / action | `seedance-2-0-image-to-video` | Nessun personaggio; qualità cinematica epica, fino a 15s |
+| Establishing / mood / action | `seedance-2-0-image-to-video-api` | Nessun personaggio; qualità cinematica epica, fino a 15s |
 
 Questi sono sovrascrivibili per progetto tramite `series.json → videoDefaults`. Per puntare a una famiglia non Seedance (es. account che non hanno accesso a Seedance), imposta `videoDefaults` su `kling-o3-standard-reference-to-video` e `veo3.1-fast-image-to-video`.
 
diff --git a/it/guides/media/seedance-2-0.mdx b/it/guides/media/seedance-2-0.mdx
index 01e08499..efea4fcb 100644
--- a/it/guides/media/seedance-2-0.mdx
+++ b/it/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/it/guides/media/video-generation.mdx b/it/guides/media/video-generation.mdx
index b5d6da5b..afb480c4 100644
--- a/it/guides/media/video-generation.mdx
+++ b/it/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Ottieni il costo esatto prima di generare. Invia solo gli input di pricing (`mod
 Il preventivo è in USD.
 
 <Info>
-**I parametri della quote sono solo input di pricing.** I campi che invii a `/video/quote` — inclusi `aspect_ratio`, `audio` e `reference_video_total_duration` — vengono usati esclusivamente per calcolare il prezzo. **Non** vengono inoltrati a `/video/queue` e non influenzano il video generato. Inviare `aspect_ratio` alla quote è accettato anche per i modelli (come `seedance-2-0-image-to-video`) che rifiutano `aspect_ratio` al momento della generazione, poiché i modelli image-to-video ricavano l'aspect ratio di output dall'immagine di input.
+**I parametri della quote sono solo input di pricing.** I campi che invii a `/video/quote` — inclusi `aspect_ratio`, `audio` e `reference_video_total_duration` — vengono usati esclusivamente per calcolare il prezzo. **Non** vengono inoltrati a `/video/queue` e non influenzano il video generato. Inviare `aspect_ratio` alla quote è accettato anche per i modelli (come `seedance-2-0-image-to-video-api`) che rifiutano `aspect_ratio` al momento della generazione, poiché i modelli image-to-video ricavano l'aspect ratio di output dall'immagine di input.
 </Info>
 
 ### Prezzi reference-to-video
diff --git a/ko/api-reference/endpoint/video/queue.mdx b/ko/api-reference/endpoint/video/queue.mdx
index c90ddc41..649804ee 100644
--- a/ko/api-reference/endpoint/video/queue.mdx
+++ b/ko/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ openapi: 'POST /video/queue'
 
 ### Seedance 2.0 & 2.5
 
-`seedance-2-0-*` 및 `seedance-2-5-*` 모델(text-to-video, image-to-video, reference-to-video, 그리고 Seedance 2.0 `-fast-*` 변형)은 [Seedance 2.0 & 2.5 가이드](/guides/media/seedance-2-0)에서 네 가지 워크플로(Reference / Edit / Extend / Stitch), 패밀리별 멀티모달 한도, faces-off API 동작, 가격 정보를 확인하세요.
+`seedance-2-0-*-api` 및 `seedance-2-5-*-api` 모델(text-to-video, image-to-video, reference-to-video, 그리고 Seedance 2.0 `-fast-*` 변형)은 [Seedance 2.0 & 2.5 가이드](/guides/media/seedance-2-0)에서 네 가지 워크플로(Reference / Edit / Extend / Stitch), 패밀리별 멀티모달 한도, faces-off API 동작, 가격 정보를 확인하세요.
 
 ### Seedance faces-off API
 
diff --git a/ko/guides/integrations/venice-video-harness.mdx b/ko/guides/integrations/venice-video-harness.mdx
index 3664067c..3120e8de 100644
--- a/ko/guides/integrations/venice-video-harness.mdx
+++ b/ko/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ curl -L -o ~/.cache/whisper.cpp/ggml-base.en.bin \
 
 | Role | Default model | When used |
 |---|---|---|
-| 캐릭터 샷(1-2 캐릭터) | `seedance-2-0-reference-to-video` | 평면 `reference_image_urls`, `@Image` 태그, 최대 15초, 네이티브 스테레오 오디오의 기본 R2V |
+| 캐릭터 샷(1-2 캐릭터) | `seedance-2-0-reference-to-video-api` | 평면 `reference_image_urls`, `@Image` 태그, 최대 15초, 네이티브 스테레오 오디오의 기본 R2V |
 | 캐릭터 샷(3+ 캐릭터) | `kling-o3-standard-reference-to-video` | 멀티 캐릭터 정체성을 위해 구조화된 `elements`로 자동 fallback |
-| 설정 / 분위기 / 액션 | `seedance-2-0-image-to-video` | 캐릭터 없음. 서사적 시네마틱 품질, 최대 15초 |
+| 설정 / 분위기 / 액션 | `seedance-2-0-image-to-video-api` | 캐릭터 없음. 서사적 시네마틱 품질, 최대 15초 |
 
 이 기본값은 `series.json → videoDefaults`로 프로젝트별로 덮어쓸 수 있습니다. Seedance가 아닌 패밀리(예: Seedance 접근 권한이 없는 계정)를 타겟팅하려면 `videoDefaults`를 `kling-o3-standard-reference-to-video`와 `veo3.1-fast-image-to-video`로 설정하세요.
 
diff --git a/ko/guides/media/seedance-2-0.mdx b/ko/guides/media/seedance-2-0.mdx
index eceb0aaa..30bf8ce6 100644
--- a/ko/guides/media/seedance-2-0.mdx
+++ b/ko/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/ko/guides/media/video-generation.mdx b/ko/guides/media/video-generation.mdx
index 7bf1c6fb..1d2e2a58 100644
--- a/ko/guides/media/video-generation.mdx
+++ b/ko/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Image-to-video 모델의 경우 `image_url`을 통해 원본 이미지를 전달
 견적은 USD 단위입니다.
 
 <Info>
-**견적 파라미터는 가격 산정 input 전용입니다.** `/video/quote`에 전송하는 필드(`aspect_ratio`, `audio`, `reference_video_total_duration` 포함)는 순전히 가격을 계산하는 데만 사용됩니다. 이 필드들은 `/video/queue`로 전달되지 **않으며** 생성되는 비디오에 영향을 주지 않습니다. quote에 `aspect_ratio`를 보내는 것은 생성 시점에 `aspect_ratio`를 거부하는 모델(예: `seedance-2-0-image-to-video`)에도 허용되는데, 이는 image-to-video 모델이 입력 이미지에서 출력 종횡비를 도출하기 때문입니다.
+**견적 파라미터는 가격 산정 input 전용입니다.** `/video/quote`에 전송하는 필드(`aspect_ratio`, `audio`, `reference_video_total_duration` 포함)는 순전히 가격을 계산하는 데만 사용됩니다. 이 필드들은 `/video/queue`로 전달되지 **않으며** 생성되는 비디오에 영향을 주지 않습니다. quote에 `aspect_ratio`를 보내는 것은 생성 시점에 `aspect_ratio`를 거부하는 모델(예: `seedance-2-0-image-to-video-api`)에도 허용되는데, 이는 image-to-video 모델이 입력 이미지에서 출력 종횡비를 도출하기 때문입니다.
 </Info>
 
 ### Reference-to-video 가격
diff --git a/pt-BR/api-reference/endpoint/video/queue.mdx b/pt-BR/api-reference/endpoint/video/queue.mdx
index c261b092..0b5a444f 100644
--- a/pt-BR/api-reference/endpoint/video/queue.mdx
+++ b/pt-BR/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Os modelos privados também retornam uma `download_url` para o vídeo finalizado
 
 ### Seedance 2.0 e 2.5
 
-Para os modelos `seedance-2-0-*` e `seedance-2-5-*` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*` do Seedance 2.0), consulte o [Guia Seedance 2.0 e 2.5](/guides/media/seedance-2-0) para o modelo de quatro fluxos (Reference / Edit / Extend / Stitch), limites multimodais por família, comportamento faces-off da API e detalhes de preços.
+Para os modelos `seedance-2-0-*-api` e `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*` do Seedance 2.0), consulte o [Guia Seedance 2.0 e 2.5](/guides/media/seedance-2-0) para o modelo de quatro fluxos (Reference / Edit / Extend / Stitch), limites multimodais por família, comportamento faces-off da API e detalhes de preços.
 
 ### API Seedance faces-off
 
diff --git a/pt-BR/guides/integrations/venice-video-harness.mdx b/pt-BR/guides/integrations/venice-video-harness.mdx
index 50eb08a8..591881ce 100644
--- a/pt-BR/guides/integrations/venice-video-harness.mdx
+++ b/pt-BR/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Os padrões do harness são opinativos porque consistência é o ponto. Roteamen
 
 | Papel | Modelo padrão | Quando usado |
 |---|---|---|
-| Tomadas de personagem (1-2 personagens) | `seedance-2-0-reference-to-video` | R2V padrão com `reference_image_urls` flat, tags `@Image`, até 15s, áudio estéreo nativo |
+| Tomadas de personagem (1-2 personagens) | `seedance-2-0-reference-to-video-api` | R2V padrão com `reference_image_urls` flat, tags `@Image`, até 15s, áudio estéreo nativo |
 | Tomadas de personagem (3+ personagens) | `kling-o3-standard-reference-to-video` | Auto-fallback com `elements` estruturados para identidade multi-personagem |
-| Estabelecimento / atmosfera / ação | `seedance-2-0-image-to-video` | Sem personagens; qualidade cinemática épica, até 15s |
+| Estabelecimento / atmosfera / ação | `seedance-2-0-image-to-video-api` | Sem personagens; qualidade cinemática épica, até 15s |
 
 Esses são sobrescritíveis por projeto via `series.json → videoDefaults`. Para mirar em uma família que não seja Seedance (por exemplo, contas que não têm acesso ao Seedance), defina `videoDefaults` como `kling-o3-standard-reference-to-video` e `veo3.1-fast-image-to-video`.
 
diff --git a/pt-BR/guides/media/seedance-2-0.mdx b/pt-BR/guides/media/seedance-2-0.mdx
index 26e4f08d..25408f5c 100644
--- a/pt-BR/guides/media/seedance-2-0.mdx
+++ b/pt-BR/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/pt-BR/guides/media/video-generation.mdx b/pt-BR/guides/media/video-generation.mdx
index b98acfdb..cbeb81ad 100644
--- a/pt-BR/guides/media/video-generation.mdx
+++ b/pt-BR/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtenha o custo exato antes de gerar. Envie apenas inputs de preço (`model`, `d
 A cotação está em USD.
 
 <Info>
-**Os parâmetros de cotação são apenas inputs de preço.** Os campos que você envia para `/video/quote` — incluindo `aspect_ratio`, `audio` e `reference_video_total_duration` — são usados exclusivamente para calcular o preço. Eles **não** são encaminhados para `/video/queue` e não afetam o vídeo gerado. Enviar `aspect_ratio` para a cotação é aceito mesmo para modelos (como `seedance-2-0-image-to-video`) que rejeitam `aspect_ratio` no momento da geração, já que modelos image-to-video derivam o aspect ratio de saída a partir da imagem de entrada.
+**Os parâmetros de cotação são apenas inputs de preço.** Os campos que você envia para `/video/quote` — incluindo `aspect_ratio`, `audio` e `reference_video_total_duration` — são usados exclusivamente para calcular o preço. Eles **não** são encaminhados para `/video/queue` e não afetam o vídeo gerado. Enviar `aspect_ratio` para a cotação é aceito mesmo para modelos (como `seedance-2-0-image-to-video-api`) que rejeitam `aspect_ratio` no momento da geração, já que modelos image-to-video derivam o aspect ratio de saída a partir da imagem de entrada.
 </Info>
 
 ### Preços de reference-to-video
diff --git a/zh/api-reference/endpoint/video/queue.mdx b/zh/api-reference/endpoint/video/queue.mdx
index a1fe02a7..0bce55da 100644
--- a/zh/api-reference/endpoint/video/queue.mdx
+++ b/zh/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ openapi: 'POST /video/queue'
 
 ### Seedance 2.0 与 2.5
 
-对于 `seedance-2-0-*` 与 `seedance-2-5-*` 模型（text-to-video、image-to-video、reference-to-video，以及 Seedance 2.0 的 `-fast-*` 变体），请参阅 [Seedance 2.0 与 2.5 指南](/guides/media/seedance-2-0)，了解四工作流模型（Reference / Edit / Extend / Stitch）、按系列区分的多模态限制、faces-off API 行为与定价说明。
+对于 `seedance-2-0-*-api` 与 `seedance-2-5-*-api` 模型（text-to-video、image-to-video、reference-to-video，以及 Seedance 2.0 的 `-fast-*` 变体），请参阅 [Seedance 2.0 与 2.5 指南](/guides/media/seedance-2-0)，了解四工作流模型（Reference / Edit / Extend / Stitch）、按系列区分的多模态限制、faces-off API 行为与定价说明。
 
 ### Seedance faces-off API
 
diff --git a/zh/guides/integrations/venice-video-harness.mdx b/zh/guides/integrations/venice-video-harness.mdx
index d72aaf9d..ebea769f 100644
--- a/zh/guides/integrations/venice-video-harness.mdx
+++ b/zh/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ harness 默认值是有主见的，因为一致性是关键。当前路由（202
 
 | 角色 | 默认模型 | 使用场景 |
 |---|---|---|
-| 角色镜头（1-2 个角色） | `seedance-2-0-reference-to-video` | 默认 R2V，采用扁平 `reference_image_urls`、`@Image` 标签，最长 15 秒，原生立体声音频 |
+| 角色镜头（1-2 个角色） | `seedance-2-0-reference-to-video-api` | 默认 R2V，采用扁平 `reference_image_urls`、`@Image` 标签，最长 15 秒，原生立体声音频 |
 | 角色镜头（3+ 个角色） | `kling-o3-standard-reference-to-video` | 自动回退，采用结构化 `elements` 处理多角色身份 |
-| 建立镜头/氛围/动作 | `seedance-2-0-image-to-video` | 无角色；史诗级电影质量，最长 15 秒 |
+| 建立镜头/氛围/动作 | `seedance-2-0-image-to-video-api` | 无角色；史诗级电影质量，最长 15 秒 |
 
 这些可以通过 `series.json → videoDefaults` 按项目覆盖。要针对非 Seedance 系列（例如缺乏 Seedance 访问权限的账户），将 `videoDefaults` 设置为 `kling-o3-standard-reference-to-video` 和 `veo3.1-fast-image-to-video`。
 
diff --git a/zh/guides/media/seedance-2-0.mdx b/zh/guides/media/seedance-2-0.mdx
index a056f013..4ed1082d 100644
--- a/zh/guides/media/seedance-2-0.mdx
+++ b/zh/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video`, its Fast sibling, and `seedance-2-5-reference-to-video`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video",
+    "model": "seedance-2-0-text-to-video-api",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video",
+    "model": "seedance-2-5-text-to-video-api",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video",
+    "model": "seedance-2-0-image-to-video-api",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video`, its Fast variant, and `seedance-2-5-image-to-video`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video",
+    "model": "seedance-2-5-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video",
+    "model": "seedance-2-0-reference-to-video-api",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/zh/guides/media/video-generation.mdx b/zh/guides/media/video-generation.mdx
index 04d1731e..96580f48 100644
--- a/zh/guides/media/video-generation.mdx
+++ b/zh/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ await fetch(`${BASE_URL}/video/complete`, {
 报价以美元为单位。
 
 <Info>
-**报价参数仅用作定价输入。** 您发送到 `/video/quote` 的字段——包括 `aspect_ratio`、`audio` 和 `reference_video_total_duration`——仅用于计算价格。它们**不会**被转发到 `/video/queue`，也不会影响生成的视频。即使某些模型（例如 `seedance-2-0-image-to-video`）在生成时会拒绝 `aspect_ratio`，向 quote 发送 `aspect_ratio` 也是可接受的，因为 image-to-video 模型会从输入图像推导出输出宽高比。
+**报价参数仅用作定价输入。** 您发送到 `/video/quote` 的字段——包括 `aspect_ratio`、`audio` 和 `reference_video_total_duration`——仅用于计算价格。它们**不会**被转发到 `/video/queue`，也不会影响生成的视频。即使某些模型（例如 `seedance-2-0-image-to-video-api`）在生成时会拒绝 `aspect_ratio`，向 quote 发送 `aspect_ratio` 也是可接受的，因为 image-to-video 模型会从输入图像推导出输出宽高比。
 </Info>
 
 ### 参考视频到视频的定价

From 687592372fc232fafc943d86dd1ef5326876e594 Mon Sep 17 00:00:00 2001
From: Tomas Bystrican <tom.bystrican@gmail.com>
Date: Mon, 10 Aug 2026 09:14:43 +0930
Subject: [PATCH 4/5] Rename documented Seedance public ids from *-api to
 *-basic (API-103).

Match ENG-950 Option 3 listed twin naming.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 api-reference/endpoint/video/queue.mdx        |  2 +-
 ar/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 ar/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 ar/guides/media/video-generation.mdx          |  2 +-
 de/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 de/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 de/guides/media/video-generation.mdx          |  2 +-
 es/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 es/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 es/guides/media/video-generation.mdx          |  2 +-
 fr/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 fr/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 fr/guides/media/video-generation.mdx          |  2 +-
 guides/integrations/venice-video-harness.mdx  |  4 +-
 guides/media/seedance-2-0.mdx                 | 48 +++++++++----------
 guides/media/video-generation.mdx             |  2 +-
 it/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 it/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 it/guides/media/video-generation.mdx          |  2 +-
 ko/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 ko/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 ko/guides/media/video-generation.mdx          |  2 +-
 pt-BR/api-reference/endpoint/video/queue.mdx  |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 pt-BR/guides/media/seedance-2-0.mdx           | 48 +++++++++----------
 pt-BR/guides/media/video-generation.mdx       |  2 +-
 zh/api-reference/endpoint/video/queue.mdx     |  2 +-
 .../integrations/venice-video-harness.mdx     |  4 +-
 zh/guides/media/seedance-2-0.mdx              | 48 +++++++++----------
 zh/guides/media/video-generation.mdx          |  2 +-
 36 files changed, 252 insertions(+), 252 deletions(-)

diff --git a/api-reference/endpoint/video/queue.mdx b/api-reference/endpoint/video/queue.mdx
index 96fd855a..e9ef252e 100644
--- a/api-reference/endpoint/video/queue.mdx
+++ b/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Private models also return a `download_url` for the finished video. It is a shor
 
 ### Seedance 2.0 & 2.5
 
-For `seedance-2-0-*-api` and `seedance-2-5-*-api` models (text-to-video, image-to-video, reference-to-video, plus the Seedance 2.0 `-fast-*` variants), see the [Seedance 2.0 & 2.5 Guide](/guides/media/seedance-2-0) for the four-workflow model (Reference / Edit / Extend / Stitch), family-specific multimodal limits, faces-off API behavior, and pricing details.
+For `seedance-2-0-*-basic` and `seedance-2-5-*-basic` models (text-to-video, image-to-video, reference-to-video, plus the Seedance 2.0 `-fast-*` variants), see the [Seedance 2.0 & 2.5 Guide](/guides/media/seedance-2-0) for the four-workflow model (Reference / Edit / Extend / Stitch), family-specific multimodal limits, faces-off API behavior, and pricing details.
 
 ### Seedance faces-off API
 
diff --git a/ar/api-reference/endpoint/video/queue.mdx b/ar/api-reference/endpoint/video/queue.mdx
index d1625e5b..c79fa617 100644
--- a/ar/api-reference/endpoint/video/queue.mdx
+++ b/ar/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ openapi: 'POST /video/queue'
 
 ### Seedance 2.0 و 2.5
 
-لنماذج `seedance-2-0-*-api` و `seedance-2-5-*-api` (text-to-video و image-to-video و reference-to-video، بالإضافة إلى متغيّرات `-fast-*` لـ Seedance 2.0)، راجع [دليل Seedance 2.0 و 2.5](/guides/media/seedance-2-0) لنموذج سير العمل الأربعة (Reference / Edit / Extend / Stitch) وحدود الوسائط المتعددة حسب العائلة وسلوك واجهة faces-off وتفاصيل التسعير.
+لنماذج `seedance-2-0-*-basic` و `seedance-2-5-*-basic` (text-to-video و image-to-video و reference-to-video، بالإضافة إلى متغيّرات `-fast-*` لـ Seedance 2.0)، راجع [دليل Seedance 2.0 و 2.5](/guides/media/seedance-2-0) لنموذج سير العمل الأربعة (Reference / Edit / Extend / Stitch) وحدود الوسائط المتعددة حسب العائلة وسلوك واجهة faces-off وتفاصيل التسعير.
 
 ### واجهة Seedance بدون وجوه (faces-off)
 
diff --git a/ar/guides/integrations/venice-video-harness.mdx b/ar/guides/integrations/venice-video-harness.mdx
index 429d326b..eec52cf4 100644
--- a/ar/guides/integrations/venice-video-harness.mdx
+++ b/ar/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ curl -L -o ~/.cache/whisper.cpp/ggml-base.en.bin \
 
 | الدور | النموذج الافتراضي | متى يُستخدم |
 |---|---|---|
-| لقطات الشخصيات (1-2 شخصية) | `seedance-2-0-reference-to-video-api` | R2V افتراضي مع `reference_image_urls` مسطّحة، ووسوم `@Image`، حتى 15 ثانية، وصوت ستيريو أصلي |
+| لقطات الشخصيات (1-2 شخصية) | `seedance-2-0-reference-to-video-basic` | R2V افتراضي مع `reference_image_urls` مسطّحة، ووسوم `@Image`، حتى 15 ثانية، وصوت ستيريو أصلي |
 | لقطات الشخصيات (3+ شخصيات) | `kling-o3-standard-reference-to-video` | بديل احتياطي تلقائي مع `elements` مهيكلة لهوية متعدّدة الشخصيات |
-| تأسيسي / مزاجي / حركة | `seedance-2-0-image-to-video-api` | بدون شخصيات؛ جودة سينمائية ملحمية، حتى 15 ثانية |
+| تأسيسي / مزاجي / حركة | `seedance-2-0-image-to-video-basic` | بدون شخصيات؛ جودة سينمائية ملحمية، حتى 15 ثانية |
 
 يمكن تجاوز هذه القيم لكل مشروع عبر `series.json → videoDefaults`. لاستهداف عائلة غير Seedance (مثل الحسابات التي تفتقر إلى وصول Seedance)، اضبط `videoDefaults` إلى `kling-o3-standard-reference-to-video` و`veo3.1-fast-image-to-video`.
 
diff --git a/ar/guides/media/seedance-2-0.mdx b/ar/guides/media/seedance-2-0.mdx
index e557ad9b..a94123a0 100644
--- a/ar/guides/media/seedance-2-0.mdx
+++ b/ar/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/ar/guides/media/video-generation.mdx b/ar/guides/media/video-generation.mdx
index a8996694..5d538e5a 100644
--- a/ar/guides/media/video-generation.mdx
+++ b/ar/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ await fetch(`${BASE_URL}/video/complete`, {
 عرض السعر بالدولار الأمريكي.
 
 <Info>
-**معاملات عرض السعر هي مدخلات تسعير فقط.** الحقول التي ترسلها إلى `/video/quote` — بما في ذلك `aspect_ratio` و`audio` و`reference_video_total_duration` — تُستخدم بحتًا لحساب السعر. **لا** تُمرَّر إلى `/video/queue` ولا تؤثر في الفيديو المولَّد. يُقبل إرسال `aspect_ratio` إلى quote حتى بالنسبة للنماذج (مثل `seedance-2-0-image-to-video-api`) التي ترفض `aspect_ratio` وقت التوليد، لأن نماذج الصورة إلى فيديو تشتق نسبة العرض إلى الارتفاع للإخراج من صورة الإدخال.
+**معاملات عرض السعر هي مدخلات تسعير فقط.** الحقول التي ترسلها إلى `/video/quote` — بما في ذلك `aspect_ratio` و`audio` و`reference_video_total_duration` — تُستخدم بحتًا لحساب السعر. **لا** تُمرَّر إلى `/video/queue` ولا تؤثر في الفيديو المولَّد. يُقبل إرسال `aspect_ratio` إلى quote حتى بالنسبة للنماذج (مثل `seedance-2-0-image-to-video-basic`) التي ترفض `aspect_ratio` وقت التوليد، لأن نماذج الصورة إلى فيديو تشتق نسبة العرض إلى الارتفاع للإخراج من صورة الإدخال.
 </Info>
 
 ### تسعير المرجع إلى فيديو
diff --git a/de/api-reference/endpoint/video/queue.mdx b/de/api-reference/endpoint/video/queue.mdx
index eb5f84ee..3594795d 100644
--- a/de/api-reference/endpoint/video/queue.mdx
+++ b/de/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Private Modelle geben außerdem eine `download_url` für das fertige Video zurü
 
 ### Seedance 2.0 & 2.5
 
-Für `seedance-2-0-*-api`- und `seedance-2-5-*-api`-Modelle (Text-to-Video, Image-to-Video, Reference-to-Video sowie die Seedance-2.0-`-fast-*`-Varianten) siehe den [Seedance-2.0-&-2.5-Guide](/guides/media/seedance-2-0) zu den vier Workflows (Reference / Edit / Extend / Stitch), familienspezifischen Multimodal-Limits, Faces-off-API-Verhalten und Preisen.
+Für `seedance-2-0-*-basic`- und `seedance-2-5-*-basic`-Modelle (Text-to-Video, Image-to-Video, Reference-to-Video sowie die Seedance-2.0-`-fast-*`-Varianten) siehe den [Seedance-2.0-&-2.5-Guide](/guides/media/seedance-2-0) zu den vier Workflows (Reference / Edit / Extend / Stitch), familienspezifischen Multimodal-Limits, Faces-off-API-Verhalten und Preisen.
 
 ### Seedance Faces-off API
 
diff --git a/de/guides/integrations/venice-video-harness.mdx b/de/guides/integrations/venice-video-harness.mdx
index 337d30b1..3e08f32a 100644
--- a/de/guides/integrations/venice-video-harness.mdx
+++ b/de/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Die Standardeinstellungen sind bewusst meinungsstark, weil Konsistenz das Ziel i
 
 | Rolle | Standardmodell | Wann verwendet |
 |---|---|---|
-| Charakter-Shots (1–2 Charaktere) | `seedance-2-0-reference-to-video-api` | Standard-R2V mit flachen `reference_image_urls`, `@Image`-Tags, bis zu 15 s, nativer Stereo-Sound |
+| Charakter-Shots (1–2 Charaktere) | `seedance-2-0-reference-to-video-basic` | Standard-R2V mit flachen `reference_image_urls`, `@Image`-Tags, bis zu 15 s, nativer Stereo-Sound |
 | Charakter-Shots (3+ Charaktere) | `kling-o3-standard-reference-to-video` | Auto-Fallback mit strukturierten `elements` für Multi-Charakter-Identität |
-| Establishing / Stimmung / Action | `seedance-2-0-image-to-video-api` | Keine Charaktere; epische Kino-Qualität, bis zu 15 s |
+| Establishing / Stimmung / Action | `seedance-2-0-image-to-video-basic` | Keine Charaktere; epische Kino-Qualität, bis zu 15 s |
 
 Diese sind pro Projekt über `series.json → videoDefaults` überschreibbar. Für eine Nicht-Seedance-Familie (z. B. Accounts ohne Seedance-Zugriff) `videoDefaults` auf `kling-o3-standard-reference-to-video` und `veo3.1-fast-image-to-video` setzen.
 
diff --git a/de/guides/media/seedance-2-0.mdx b/de/guides/media/seedance-2-0.mdx
index 8443a526..f1ef7bdd 100644
--- a/de/guides/media/seedance-2-0.mdx
+++ b/de/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/de/guides/media/video-generation.mdx b/de/guides/media/video-generation.mdx
index 7d94d89b..f494cb4b 100644
--- a/de/guides/media/video-generation.mdx
+++ b/de/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Exakte Kosten vor der Generierung. Nur die Pricing-Inputs senden (`model`, `dura
 Das Quote ist in USD.
 
 <Info>
-**Quote-Parameter sind ausschließlich Preis-Inputs.** Felder, die Sie an `/video/quote` senden – darunter `aspect_ratio`, `audio` und `reference_video_total_duration` – dienen ausschließlich der Preisberechnung. Sie werden **nicht** an `/video/queue` weitergeleitet und haben keinen Einfluss auf das generierte Video. `aspect_ratio` an Quote zu senden ist auch bei Modellen zulässig (z. B. `seedance-2-0-image-to-video-api`), die `aspect_ratio` bei der Generierung ablehnen, da Image-to-Video-Modelle das Ausgabe-Seitenverhältnis aus dem Eingabebild ableiten.
+**Quote-Parameter sind ausschließlich Preis-Inputs.** Felder, die Sie an `/video/quote` senden – darunter `aspect_ratio`, `audio` und `reference_video_total_duration` – dienen ausschließlich der Preisberechnung. Sie werden **nicht** an `/video/queue` weitergeleitet und haben keinen Einfluss auf das generierte Video. `aspect_ratio` an Quote zu senden ist auch bei Modellen zulässig (z. B. `seedance-2-0-image-to-video-basic`), die `aspect_ratio` bei der Generierung ablehnen, da Image-to-Video-Modelle das Ausgabe-Seitenverhältnis aus dem Eingabebild ableiten.
 </Info>
 
 ### Preisberechnung bei Reference-to-Video
diff --git a/es/api-reference/endpoint/video/queue.mdx b/es/api-reference/endpoint/video/queue.mdx
index f5035f23..92410eba 100644
--- a/es/api-reference/endpoint/video/queue.mdx
+++ b/es/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Los modelos privados también devuelven un `download_url` para el vídeo termina
 
 ### Seedance 2.0 y 2.5
 
-Para los modelos `seedance-2-0-*-api` y `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*` de Seedance 2.0), consulta la [Guía de Seedance 2.0 y 2.5](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), límites multimodales por familia, comportamiento faces-off de la API y detalles de precios.
+Para los modelos `seedance-2-0-*-basic` y `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*` de Seedance 2.0), consulta la [Guía de Seedance 2.0 y 2.5](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), límites multimodales por familia, comportamiento faces-off de la API y detalles de precios.
 
 ### Seedance API sin rostros (faces-off)
 
diff --git a/es/guides/integrations/venice-video-harness.mdx b/es/guides/integrations/venice-video-harness.mdx
index e06c82c2..76d1b485 100644
--- a/es/guides/integrations/venice-video-harness.mdx
+++ b/es/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Los defaults del harness son opinionados porque la consistencia es lo importante
 
 | Rol | Modelo predeterminado | Cuándo se usa |
 |---|---|---|
-| Tomas de personaje (1-2 personajes) | `seedance-2-0-reference-to-video-api` | R2V por defecto con `reference_image_urls` planos, etiquetas `@Image`, hasta 15s, audio estéreo nativo |
+| Tomas de personaje (1-2 personajes) | `seedance-2-0-reference-to-video-basic` | R2V por defecto con `reference_image_urls` planos, etiquetas `@Image`, hasta 15s, audio estéreo nativo |
 | Tomas de personaje (3+ personajes) | `kling-o3-standard-reference-to-video` | Fallback automático con `elements` estructurados para identidad multi-personaje |
-| Establecimiento / atmósfera / acción | `seedance-2-0-image-to-video-api` | Sin personajes; calidad cinematográfica épica, hasta 15s |
+| Establecimiento / atmósfera / acción | `seedance-2-0-image-to-video-basic` | Sin personajes; calidad cinematográfica épica, hasta 15s |
 
 Estos son sobrescribibles por proyecto vía `series.json → videoDefaults`. Para apuntar a una familia que no sea Seedance (p. ej., cuentas sin acceso a Seedance), establece `videoDefaults` en `kling-o3-standard-reference-to-video` y `veo3.1-fast-image-to-video`.
 
diff --git a/es/guides/media/seedance-2-0.mdx b/es/guides/media/seedance-2-0.mdx
index 01daca33..d202b63c 100644
--- a/es/guides/media/seedance-2-0.mdx
+++ b/es/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/es/guides/media/video-generation.mdx b/es/guides/media/video-generation.mdx
index 2ec1e380..31d9d880 100644
--- a/es/guides/media/video-generation.mdx
+++ b/es/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtén el coste exacto antes de generar. Envía solo entradas de precio (`model`
 La cotización es en USD.
 
 <Info>
-**Los parámetros de cotización son solo entradas de precio.** Los campos que envías a `/video/quote` —incluyendo `aspect_ratio`, `audio` y `reference_video_total_duration`— se usan únicamente para calcular el precio. **No** se reenvían a `/video/queue` y no afectan al vídeo generado. Enviar `aspect_ratio` a quote se acepta incluso para modelos (como `seedance-2-0-image-to-video-api`) que rechazan `aspect_ratio` en el momento de la generación, ya que los modelos image-to-video derivan la relación de aspecto de salida a partir de la imagen de entrada.
+**Los parámetros de cotización son solo entradas de precio.** Los campos que envías a `/video/quote` —incluyendo `aspect_ratio`, `audio` y `reference_video_total_duration`— se usan únicamente para calcular el precio. **No** se reenvían a `/video/queue` y no afectan al vídeo generado. Enviar `aspect_ratio` a quote se acepta incluso para modelos (como `seedance-2-0-image-to-video-basic`) que rechazan `aspect_ratio` en el momento de la generación, ya que los modelos image-to-video derivan la relación de aspecto de salida a partir de la imagen de entrada.
 </Info>
 
 ### Precios de reference-to-video
diff --git a/fr/api-reference/endpoint/video/queue.mdx b/fr/api-reference/endpoint/video/queue.mdx
index b6bada21..bae0b72b 100644
--- a/fr/api-reference/endpoint/video/queue.mdx
+++ b/fr/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Les modèles privés renvoient également un `download_url` pour la vidéo termi
 
 ### Seedance 2.0 et 2.5
 
-Pour les modèles `seedance-2-0-*-api` et `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*` de Seedance 2.0), voir le [Guide Seedance 2.0 et 2.5](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les limites multimodales par famille, le comportement faces-off de l’API et les tarifs.
+Pour les modèles `seedance-2-0-*-basic` et `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*` de Seedance 2.0), voir le [Guide Seedance 2.0 et 2.5](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les limites multimodales par famille, le comportement faces-off de l’API et les tarifs.
 
 ### API Seedance faces-off
 
diff --git a/fr/guides/integrations/venice-video-harness.mdx b/fr/guides/integrations/venice-video-harness.mdx
index 5399da4c..ff88276f 100644
--- a/fr/guides/integrations/venice-video-harness.mdx
+++ b/fr/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Les défauts du harness sont assumés parce que la cohérence est le but. Le rou
 
 | Rôle | Modèle par défaut | Quand utilisé |
 |---|---|---|
-| Plans de personnage (1-2 personnages) | `seedance-2-0-reference-to-video-api` | R2V par défaut avec `reference_image_urls` plats, tags `@Image`, jusqu'à 15 s, audio stéréo natif |
+| Plans de personnage (1-2 personnages) | `seedance-2-0-reference-to-video-basic` | R2V par défaut avec `reference_image_urls` plats, tags `@Image`, jusqu'à 15 s, audio stéréo natif |
 | Plans de personnage (3+ personnages) | `kling-o3-standard-reference-to-video` | Fallback automatique avec `elements` structurés pour l'identité multi-personnages |
-| Plan d'établissement / d'ambiance / d'action | `seedance-2-0-image-to-video-api` | Pas de personnages ; qualité cinématique épique, jusqu'à 15 s |
+| Plan d'établissement / d'ambiance / d'action | `seedance-2-0-image-to-video-basic` | Pas de personnages ; qualité cinématique épique, jusqu'à 15 s |
 
 Ces valeurs sont remplaçables par projet via `series.json → videoDefaults`. Pour cibler une famille non-Seedance (par ex. des comptes qui n'ont pas accès à Seedance), définissez `videoDefaults` à `kling-o3-standard-reference-to-video` et `veo3.1-fast-image-to-video`.
 
diff --git a/fr/guides/media/seedance-2-0.mdx b/fr/guides/media/seedance-2-0.mdx
index 683d51a4..e88d52d1 100644
--- a/fr/guides/media/seedance-2-0.mdx
+++ b/fr/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/fr/guides/media/video-generation.mdx b/fr/guides/media/video-generation.mdx
index 141cf9ae..5474cad8 100644
--- a/fr/guides/media/video-generation.mdx
+++ b/fr/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtenez le coût exact avant la génération. Envoyez uniquement les entrées de
 Le devis est en USD.
 
 <Info>
-**Les paramètres du devis servent uniquement au calcul du prix.** Les champs que vous envoyez à `/video/quote` — y compris `aspect_ratio`, `audio` et `reference_video_total_duration` — sont utilisés uniquement pour calculer le prix. Ils **ne sont pas** transmis à `/video/queue` et n'influencent pas la vidéo générée. Envoyer `aspect_ratio` au quote est accepté même pour les modèles (tels que `seedance-2-0-image-to-video-api`) qui rejettent `aspect_ratio` au moment de la génération, puisque les modèles image-to-video dérivent le ratio d'aspect de sortie à partir de l'image d'entrée.
+**Les paramètres du devis servent uniquement au calcul du prix.** Les champs que vous envoyez à `/video/quote` — y compris `aspect_ratio`, `audio` et `reference_video_total_duration` — sont utilisés uniquement pour calculer le prix. Ils **ne sont pas** transmis à `/video/queue` et n'influencent pas la vidéo générée. Envoyer `aspect_ratio` au quote est accepté même pour les modèles (tels que `seedance-2-0-image-to-video-basic`) qui rejettent `aspect_ratio` au moment de la génération, puisque les modèles image-to-video dérivent le ratio d'aspect de sortie à partir de l'image d'entrée.
 </Info>
 
 ### Tarification reference-to-video
diff --git a/guides/integrations/venice-video-harness.mdx b/guides/integrations/venice-video-harness.mdx
index 15dce8a1..1b318fe1 100644
--- a/guides/integrations/venice-video-harness.mdx
+++ b/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ The harness defaults are opinionated because consistency is the point. The curre
 
 | Role | Default model | When used |
 |---|---|---|
-| Character shots (1-2 characters) | `seedance-2-0-reference-to-video-api` | Default R2V with flat `reference_image_urls`, `@Image` tags, up to 15s, native stereo audio |
+| Character shots (1-2 characters) | `seedance-2-0-reference-to-video-basic` | Default R2V with flat `reference_image_urls`, `@Image` tags, up to 15s, native stereo audio |
 | Character shots (3+ characters) | `kling-o3-standard-reference-to-video` | Auto-fallback with structured `elements` for multi-character identity |
-| Establishing / mood / action | `seedance-2-0-image-to-video-api` | No characters; epic cinematic quality, up to 15s |
+| Establishing / mood / action | `seedance-2-0-image-to-video-basic` | No characters; epic cinematic quality, up to 15s |
 
 These are overridable per-project via `series.json → videoDefaults`. To target a non-Seedance family (e.g. accounts that lack Seedance access), set `videoDefaults` to `kling-o3-standard-reference-to-video` and `veo3.1-fast-image-to-video`.
 
diff --git a/guides/media/seedance-2-0.mdx b/guides/media/seedance-2-0.mdx
index ef3f1337..ba46a9d6 100644
--- a/guides/media/seedance-2-0.mdx
+++ b/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/guides/media/video-generation.mdx b/guides/media/video-generation.mdx
index 72d0084d..19514f3a 100644
--- a/guides/media/video-generation.mdx
+++ b/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Get exact cost before generating. Send only pricing inputs (`model`, `duration`,
 Quote is in USD.
 
 <Info>
-**Quote parameters are pricing inputs only.** Fields you send to `/video/quote` — including `aspect_ratio`, `audio`, and `reference_video_total_duration` — are used purely to compute the price. They are **not** forwarded to `/video/queue` and do not affect the generated video. Sending `aspect_ratio` to quote is accepted even for models (such as `seedance-2-0-image-to-video-api`) that reject `aspect_ratio` at generation time, since image-to-video models derive the output aspect ratio from the input image.
+**Quote parameters are pricing inputs only.** Fields you send to `/video/quote` — including `aspect_ratio`, `audio`, and `reference_video_total_duration` — are used purely to compute the price. They are **not** forwarded to `/video/queue` and do not affect the generated video. Sending `aspect_ratio` to quote is accepted even for models (such as `seedance-2-0-image-to-video-basic`) that reject `aspect_ratio` at generation time, since image-to-video models derive the output aspect ratio from the input image.
 </Info>
 
 ### Reference-to-video pricing
diff --git a/it/api-reference/endpoint/video/queue.mdx b/it/api-reference/endpoint/video/queue.mdx
index cb62b18b..b3fc9278 100644
--- a/it/api-reference/endpoint/video/queue.mdx
+++ b/it/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ I modelli privati restituiscono anche un `download_url` per il video finito. Si
 
 ### Seedance 2.0 e 2.5
 
-Per i modelli `seedance-2-0-*-api` e `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*` di Seedance 2.0), vedi la [Guida Seedance 2.0 e 2.5](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i limiti multimodali per famiglia, il comportamento faces-off dell’API e i dettagli sui prezzi.
+Per i modelli `seedance-2-0-*-basic` e `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*` di Seedance 2.0), vedi la [Guida Seedance 2.0 e 2.5](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i limiti multimodali per famiglia, il comportamento faces-off dell’API e i dettagli sui prezzi.
 
 ### API Seedance faces-off
 
diff --git a/it/guides/integrations/venice-video-harness.mdx b/it/guides/integrations/venice-video-harness.mdx
index 9684ee68..7a69efb6 100644
--- a/it/guides/integrations/venice-video-harness.mdx
+++ b/it/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ I default dell'harness sono opinionated perché la consistenza è il punto chiav
 
 | Ruolo | Modello di default | Quando usato |
 |---|---|---|
-| Inquadrature con personaggi (1-2 personaggi) | `seedance-2-0-reference-to-video-api` | R2V di default con `reference_image_urls` flat, tag `@Image`, fino a 15s, audio stereo nativo |
+| Inquadrature con personaggi (1-2 personaggi) | `seedance-2-0-reference-to-video-basic` | R2V di default con `reference_image_urls` flat, tag `@Image`, fino a 15s, audio stereo nativo |
 | Inquadrature con personaggi (3+ personaggi) | `kling-o3-standard-reference-to-video` | Fallback automatico con `elements` strutturati per identità multi-personaggio |
-| Establishing / mood / action | `seedance-2-0-image-to-video-api` | Nessun personaggio; qualità cinematica epica, fino a 15s |
+| Establishing / mood / action | `seedance-2-0-image-to-video-basic` | Nessun personaggio; qualità cinematica epica, fino a 15s |
 
 Questi sono sovrascrivibili per progetto tramite `series.json → videoDefaults`. Per puntare a una famiglia non Seedance (es. account che non hanno accesso a Seedance), imposta `videoDefaults` su `kling-o3-standard-reference-to-video` e `veo3.1-fast-image-to-video`.
 
diff --git a/it/guides/media/seedance-2-0.mdx b/it/guides/media/seedance-2-0.mdx
index efea4fcb..947767b5 100644
--- a/it/guides/media/seedance-2-0.mdx
+++ b/it/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/it/guides/media/video-generation.mdx b/it/guides/media/video-generation.mdx
index afb480c4..b820d7b1 100644
--- a/it/guides/media/video-generation.mdx
+++ b/it/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Ottieni il costo esatto prima di generare. Invia solo gli input di pricing (`mod
 Il preventivo è in USD.
 
 <Info>
-**I parametri della quote sono solo input di pricing.** I campi che invii a `/video/quote` — inclusi `aspect_ratio`, `audio` e `reference_video_total_duration` — vengono usati esclusivamente per calcolare il prezzo. **Non** vengono inoltrati a `/video/queue` e non influenzano il video generato. Inviare `aspect_ratio` alla quote è accettato anche per i modelli (come `seedance-2-0-image-to-video-api`) che rifiutano `aspect_ratio` al momento della generazione, poiché i modelli image-to-video ricavano l'aspect ratio di output dall'immagine di input.
+**I parametri della quote sono solo input di pricing.** I campi che invii a `/video/quote` — inclusi `aspect_ratio`, `audio` e `reference_video_total_duration` — vengono usati esclusivamente per calcolare il prezzo. **Non** vengono inoltrati a `/video/queue` e non influenzano il video generato. Inviare `aspect_ratio` alla quote è accettato anche per i modelli (come `seedance-2-0-image-to-video-basic`) che rifiutano `aspect_ratio` al momento della generazione, poiché i modelli image-to-video ricavano l'aspect ratio di output dall'immagine di input.
 </Info>
 
 ### Prezzi reference-to-video
diff --git a/ko/api-reference/endpoint/video/queue.mdx b/ko/api-reference/endpoint/video/queue.mdx
index 649804ee..03007670 100644
--- a/ko/api-reference/endpoint/video/queue.mdx
+++ b/ko/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ openapi: 'POST /video/queue'
 
 ### Seedance 2.0 & 2.5
 
-`seedance-2-0-*-api` 및 `seedance-2-5-*-api` 모델(text-to-video, image-to-video, reference-to-video, 그리고 Seedance 2.0 `-fast-*` 변형)은 [Seedance 2.0 & 2.5 가이드](/guides/media/seedance-2-0)에서 네 가지 워크플로(Reference / Edit / Extend / Stitch), 패밀리별 멀티모달 한도, faces-off API 동작, 가격 정보를 확인하세요.
+`seedance-2-0-*-basic` 및 `seedance-2-5-*-basic` 모델(text-to-video, image-to-video, reference-to-video, 그리고 Seedance 2.0 `-fast-*` 변형)은 [Seedance 2.0 & 2.5 가이드](/guides/media/seedance-2-0)에서 네 가지 워크플로(Reference / Edit / Extend / Stitch), 패밀리별 멀티모달 한도, faces-off API 동작, 가격 정보를 확인하세요.
 
 ### Seedance faces-off API
 
diff --git a/ko/guides/integrations/venice-video-harness.mdx b/ko/guides/integrations/venice-video-harness.mdx
index 3120e8de..7e50d081 100644
--- a/ko/guides/integrations/venice-video-harness.mdx
+++ b/ko/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ curl -L -o ~/.cache/whisper.cpp/ggml-base.en.bin \
 
 | Role | Default model | When used |
 |---|---|---|
-| 캐릭터 샷(1-2 캐릭터) | `seedance-2-0-reference-to-video-api` | 평면 `reference_image_urls`, `@Image` 태그, 최대 15초, 네이티브 스테레오 오디오의 기본 R2V |
+| 캐릭터 샷(1-2 캐릭터) | `seedance-2-0-reference-to-video-basic` | 평면 `reference_image_urls`, `@Image` 태그, 최대 15초, 네이티브 스테레오 오디오의 기본 R2V |
 | 캐릭터 샷(3+ 캐릭터) | `kling-o3-standard-reference-to-video` | 멀티 캐릭터 정체성을 위해 구조화된 `elements`로 자동 fallback |
-| 설정 / 분위기 / 액션 | `seedance-2-0-image-to-video-api` | 캐릭터 없음. 서사적 시네마틱 품질, 최대 15초 |
+| 설정 / 분위기 / 액션 | `seedance-2-0-image-to-video-basic` | 캐릭터 없음. 서사적 시네마틱 품질, 최대 15초 |
 
 이 기본값은 `series.json → videoDefaults`로 프로젝트별로 덮어쓸 수 있습니다. Seedance가 아닌 패밀리(예: Seedance 접근 권한이 없는 계정)를 타겟팅하려면 `videoDefaults`를 `kling-o3-standard-reference-to-video`와 `veo3.1-fast-image-to-video`로 설정하세요.
 
diff --git a/ko/guides/media/seedance-2-0.mdx b/ko/guides/media/seedance-2-0.mdx
index 30bf8ce6..c3c88435 100644
--- a/ko/guides/media/seedance-2-0.mdx
+++ b/ko/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/ko/guides/media/video-generation.mdx b/ko/guides/media/video-generation.mdx
index 1d2e2a58..3e40c6e7 100644
--- a/ko/guides/media/video-generation.mdx
+++ b/ko/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Image-to-video 모델의 경우 `image_url`을 통해 원본 이미지를 전달
 견적은 USD 단위입니다.
 
 <Info>
-**견적 파라미터는 가격 산정 input 전용입니다.** `/video/quote`에 전송하는 필드(`aspect_ratio`, `audio`, `reference_video_total_duration` 포함)는 순전히 가격을 계산하는 데만 사용됩니다. 이 필드들은 `/video/queue`로 전달되지 **않으며** 생성되는 비디오에 영향을 주지 않습니다. quote에 `aspect_ratio`를 보내는 것은 생성 시점에 `aspect_ratio`를 거부하는 모델(예: `seedance-2-0-image-to-video-api`)에도 허용되는데, 이는 image-to-video 모델이 입력 이미지에서 출력 종횡비를 도출하기 때문입니다.
+**견적 파라미터는 가격 산정 input 전용입니다.** `/video/quote`에 전송하는 필드(`aspect_ratio`, `audio`, `reference_video_total_duration` 포함)는 순전히 가격을 계산하는 데만 사용됩니다. 이 필드들은 `/video/queue`로 전달되지 **않으며** 생성되는 비디오에 영향을 주지 않습니다. quote에 `aspect_ratio`를 보내는 것은 생성 시점에 `aspect_ratio`를 거부하는 모델(예: `seedance-2-0-image-to-video-basic`)에도 허용되는데, 이는 image-to-video 모델이 입력 이미지에서 출력 종횡비를 도출하기 때문입니다.
 </Info>
 
 ### Reference-to-video 가격
diff --git a/pt-BR/api-reference/endpoint/video/queue.mdx b/pt-BR/api-reference/endpoint/video/queue.mdx
index 0b5a444f..dd58d2fa 100644
--- a/pt-BR/api-reference/endpoint/video/queue.mdx
+++ b/pt-BR/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ Os modelos privados também retornam uma `download_url` para o vídeo finalizado
 
 ### Seedance 2.0 e 2.5
 
-Para os modelos `seedance-2-0-*-api` e `seedance-2-5-*-api` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*` do Seedance 2.0), consulte o [Guia Seedance 2.0 e 2.5](/guides/media/seedance-2-0) para o modelo de quatro fluxos (Reference / Edit / Extend / Stitch), limites multimodais por família, comportamento faces-off da API e detalhes de preços.
+Para os modelos `seedance-2-0-*-basic` e `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*` do Seedance 2.0), consulte o [Guia Seedance 2.0 e 2.5](/guides/media/seedance-2-0) para o modelo de quatro fluxos (Reference / Edit / Extend / Stitch), limites multimodais por família, comportamento faces-off da API e detalhes de preços.
 
 ### API Seedance faces-off
 
diff --git a/pt-BR/guides/integrations/venice-video-harness.mdx b/pt-BR/guides/integrations/venice-video-harness.mdx
index 591881ce..7b2db68e 100644
--- a/pt-BR/guides/integrations/venice-video-harness.mdx
+++ b/pt-BR/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Os padrões do harness são opinativos porque consistência é o ponto. Roteamen
 
 | Papel | Modelo padrão | Quando usado |
 |---|---|---|
-| Tomadas de personagem (1-2 personagens) | `seedance-2-0-reference-to-video-api` | R2V padrão com `reference_image_urls` flat, tags `@Image`, até 15s, áudio estéreo nativo |
+| Tomadas de personagem (1-2 personagens) | `seedance-2-0-reference-to-video-basic` | R2V padrão com `reference_image_urls` flat, tags `@Image`, até 15s, áudio estéreo nativo |
 | Tomadas de personagem (3+ personagens) | `kling-o3-standard-reference-to-video` | Auto-fallback com `elements` estruturados para identidade multi-personagem |
-| Estabelecimento / atmosfera / ação | `seedance-2-0-image-to-video-api` | Sem personagens; qualidade cinemática épica, até 15s |
+| Estabelecimento / atmosfera / ação | `seedance-2-0-image-to-video-basic` | Sem personagens; qualidade cinemática épica, até 15s |
 
 Esses são sobrescritíveis por projeto via `series.json → videoDefaults`. Para mirar em uma família que não seja Seedance (por exemplo, contas que não têm acesso ao Seedance), defina `videoDefaults` como `kling-o3-standard-reference-to-video` e `veo3.1-fast-image-to-video`.
 
diff --git a/pt-BR/guides/media/seedance-2-0.mdx b/pt-BR/guides/media/seedance-2-0.mdx
index 25408f5c..be4cbc87 100644
--- a/pt-BR/guides/media/seedance-2-0.mdx
+++ b/pt-BR/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/pt-BR/guides/media/video-generation.mdx b/pt-BR/guides/media/video-generation.mdx
index cbeb81ad..a7c944e4 100644
--- a/pt-BR/guides/media/video-generation.mdx
+++ b/pt-BR/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtenha o custo exato antes de gerar. Envie apenas inputs de preço (`model`, `d
 A cotação está em USD.
 
 <Info>
-**Os parâmetros de cotação são apenas inputs de preço.** Os campos que você envia para `/video/quote` — incluindo `aspect_ratio`, `audio` e `reference_video_total_duration` — são usados exclusivamente para calcular o preço. Eles **não** são encaminhados para `/video/queue` e não afetam o vídeo gerado. Enviar `aspect_ratio` para a cotação é aceito mesmo para modelos (como `seedance-2-0-image-to-video-api`) que rejeitam `aspect_ratio` no momento da geração, já que modelos image-to-video derivam o aspect ratio de saída a partir da imagem de entrada.
+**Os parâmetros de cotação são apenas inputs de preço.** Os campos que você envia para `/video/quote` — incluindo `aspect_ratio`, `audio` e `reference_video_total_duration` — são usados exclusivamente para calcular o preço. Eles **não** são encaminhados para `/video/queue` e não afetam o vídeo gerado. Enviar `aspect_ratio` para a cotação é aceito mesmo para modelos (como `seedance-2-0-image-to-video-basic`) que rejeitam `aspect_ratio` no momento da geração, já que modelos image-to-video derivam o aspect ratio de saída a partir da imagem de entrada.
 </Info>
 
 ### Preços de reference-to-video
diff --git a/zh/api-reference/endpoint/video/queue.mdx b/zh/api-reference/endpoint/video/queue.mdx
index 0bce55da..5b01b9e0 100644
--- a/zh/api-reference/endpoint/video/queue.mdx
+++ b/zh/api-reference/endpoint/video/queue.mdx
@@ -11,7 +11,7 @@ openapi: 'POST /video/queue'
 
 ### Seedance 2.0 与 2.5
 
-对于 `seedance-2-0-*-api` 与 `seedance-2-5-*-api` 模型（text-to-video、image-to-video、reference-to-video，以及 Seedance 2.0 的 `-fast-*` 变体），请参阅 [Seedance 2.0 与 2.5 指南](/guides/media/seedance-2-0)，了解四工作流模型（Reference / Edit / Extend / Stitch）、按系列区分的多模态限制、faces-off API 行为与定价说明。
+对于 `seedance-2-0-*-basic` 与 `seedance-2-5-*-basic` 模型（text-to-video、image-to-video、reference-to-video，以及 Seedance 2.0 的 `-fast-*` 变体），请参阅 [Seedance 2.0 与 2.5 指南](/guides/media/seedance-2-0)，了解四工作流模型（Reference / Edit / Extend / Stitch）、按系列区分的多模态限制、faces-off API 行为与定价说明。
 
 ### Seedance faces-off API
 
diff --git a/zh/guides/integrations/venice-video-harness.mdx b/zh/guides/integrations/venice-video-harness.mdx
index ebea769f..877280e0 100644
--- a/zh/guides/integrations/venice-video-harness.mdx
+++ b/zh/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ harness 默认值是有主见的，因为一致性是关键。当前路由（202
 
 | 角色 | 默认模型 | 使用场景 |
 |---|---|---|
-| 角色镜头（1-2 个角色） | `seedance-2-0-reference-to-video-api` | 默认 R2V，采用扁平 `reference_image_urls`、`@Image` 标签，最长 15 秒，原生立体声音频 |
+| 角色镜头（1-2 个角色） | `seedance-2-0-reference-to-video-basic` | 默认 R2V，采用扁平 `reference_image_urls`、`@Image` 标签，最长 15 秒，原生立体声音频 |
 | 角色镜头（3+ 个角色） | `kling-o3-standard-reference-to-video` | 自动回退，采用结构化 `elements` 处理多角色身份 |
-| 建立镜头/氛围/动作 | `seedance-2-0-image-to-video-api` | 无角色；史诗级电影质量，最长 15 秒 |
+| 建立镜头/氛围/动作 | `seedance-2-0-image-to-video-basic` | 无角色；史诗级电影质量，最长 15 秒 |
 
 这些可以通过 `series.json → videoDefaults` 按项目覆盖。要针对非 Seedance 系列（例如缺乏 Seedance 访问权限的账户），将 `videoDefaults` 设置为 `kling-o3-standard-reference-to-video` 和 `veo3.1-fast-image-to-video`。
 
diff --git a/zh/guides/media/seedance-2-0.mdx b/zh/guides/media/seedance-2-0.mdx
index 4ed1082d..680289fc 100644
--- a/zh/guides/media/seedance-2-0.mdx
+++ b/zh/guides/media/seedance-2-0.mdx
@@ -17,15 +17,15 @@ This guide covers variants, the four workflows, **faces-off public API behavior*
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-api` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-api` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-api` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-api` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-api` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-api` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-api` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-api` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-api` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
+| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
+| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
+| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
+| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
+| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
+| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
 
 All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
 
@@ -33,7 +33,7 @@ Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seeda
 
 ## The "one model, four workflows" model
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-api`, its Fast sibling, and `seedance-2-5-reference-to-video-api`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
@@ -218,7 +218,7 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -239,7 +239,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -254,7 +254,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-api",
+    "model": "seedance-2-0-text-to-video-basic",
     "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -269,7 +269,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-text-to-video-api",
+    "model": "seedance-2-5-text-to-video-basic",
     "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
     "duration": "20s",
     "aspect_ratio": "16:9",
@@ -284,7 +284,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-api",
+    "model": "seedance-2-0-image-to-video-basic",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,7 +293,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-api`, its Fast variant, and `seedance-2-5-image-to-video-api`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
 </Note>
 
 ### Reference workflow — subject donor
@@ -303,7 +303,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -319,7 +319,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-5-reference-to-video-api",
+    "model": "seedance-2-5-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
     "reference_image_urls": [
       "https://example.com/character.png",
@@ -338,7 +338,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -355,7 +355,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -372,7 +372,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +390,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +407,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -430,7 +430,7 @@ curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-api",
+    "model": "seedance-2-0-reference-to-video-basic",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
diff --git a/zh/guides/media/video-generation.mdx b/zh/guides/media/video-generation.mdx
index 96580f48..e5577571 100644
--- a/zh/guides/media/video-generation.mdx
+++ b/zh/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ await fetch(`${BASE_URL}/video/complete`, {
 报价以美元为单位。
 
 <Info>
-**报价参数仅用作定价输入。** 您发送到 `/video/quote` 的字段——包括 `aspect_ratio`、`audio` 和 `reference_video_total_duration`——仅用于计算价格。它们**不会**被转发到 `/video/queue`，也不会影响生成的视频。即使某些模型（例如 `seedance-2-0-image-to-video-api`）在生成时会拒绝 `aspect_ratio`，向 quote 发送 `aspect_ratio` 也是可接受的，因为 image-to-video 模型会从输入图像推导出输出宽高比。
+**报价参数仅用作定价输入。** 您发送到 `/video/quote` 的字段——包括 `aspect_ratio`、`audio` 和 `reference_video_total_duration`——仅用于计算价格。它们**不会**被转发到 `/video/queue`，也不会影响生成的视频。即使某些模型（例如 `seedance-2-0-image-to-video-basic`）在生成时会拒绝 `aspect_ratio`，向 quote 发送 `aspect_ratio` 也是可接受的，因为 image-to-video 模型会从输入图像推导出输出宽高比。
 </Info>
 
 ### 参考视频到视频的定价

From bdf41357923be3b6f462dabf452335c8a9ec4a0d Mon Sep 17 00:00:00 2001
From: Tomas Bystrican <tom.bystrican@gmail.com>
Date: Tue, 11 Aug 2026 09:34:20 +0930
Subject: [PATCH 5/5] Address review: drop plan files and revert locale edits.

Keep English Seedance docs changes; let Mintlify regenerate
translations after merge.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 ar/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 ar/guides/media/seedance-2-0.mdx              | 339 +++++++----------
 ar/guides/media/video-generation.mdx          |   2 +-
 de/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 de/guides/media/seedance-2-0.mdx              | 359 ++++++++----------
 de/guides/media/video-generation.mdx          |   2 +-
 ...6-08-08-api-103-seedance-faces-off-docs.md |  25 --
 ...-api-103-seedance-faces-off-docs-design.md |  79 ----
 es/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 es/guides/media/seedance-2-0.mdx              | 343 +++++++----------
 es/guides/media/video-generation.mdx          |   2 +-
 fr/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 fr/guides/media/seedance-2-0.mdx              | 335 +++++++---------
 fr/guides/media/video-generation.mdx          |   2 +-
 it/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 it/guides/media/seedance-2-0.mdx              | 347 +++++++----------
 it/guides/media/video-generation.mdx          |   2 +-
 ko/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 ko/guides/media/seedance-2-0.mdx              | 343 +++++++----------
 ko/guides/media/video-generation.mdx          |   2 +-
 pt-BR/api-reference/endpoint/video/queue.mdx  |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 pt-BR/guides/media/seedance-2-0.mdx           | 337 +++++++---------
 pt-BR/guides/media/video-generation.mdx       |   2 +-
 zh/api-reference/endpoint/video/queue.mdx     |   8 +-
 .../integrations/venice-video-harness.mdx     |   4 +-
 zh/guides/media/seedance-2-0.mdx              | 337 +++++++---------
 zh/guides/media/video-generation.mdx          |   2 +-
 34 files changed, 1174 insertions(+), 1782 deletions(-)
 delete mode 100644 docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md
 delete mode 100644 docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md

diff --git a/ar/api-reference/endpoint/video/queue.mdx b/ar/api-reference/endpoint/video/queue.mdx
index c79fa617..0b982a72 100644
--- a/ar/api-reference/endpoint/video/queue.mdx
+++ b/ar/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ openapi: 'POST /video/queue'
 
 تُعيد النماذج الخاصة أيضًا `download_url` للفيديو النهائي. وهو عنوان URL للتسليم قصير الأمد (لا بأس بإعادة المحاولة بضع مرات في حال انقطع التنزيل)؛ راجع [دليل توليد الفيديو](/guides/media/video-generation#private-download-links) للتفاصيل و`DELETE` الاختيارية للخصوصية.
 
-### Seedance 2.0 و 2.5
+### Seedance 2.0
 
-لنماذج `seedance-2-0-*-basic` و `seedance-2-5-*-basic` (text-to-video و image-to-video و reference-to-video، بالإضافة إلى متغيّرات `-fast-*` لـ Seedance 2.0)، راجع [دليل Seedance 2.0 و 2.5](/guides/media/seedance-2-0) لنموذج سير العمل الأربعة (Reference / Edit / Extend / Stitch) وحدود الوسائط المتعددة حسب العائلة وسلوك واجهة faces-off وتفاصيل التسعير.
+بالنسبة لنماذج `seedance-2-0-*` (text-to-video، image-to-video، reference-to-video، بالإضافة إلى متغيرات `-fast-*`)، راجع [دليل Seedance 2.0](/guides/media/seedance-2-0) لنموذج التدفقات الأربعة (Reference / Edit / Extend / Stitch)، وأنماط الـ prompt القياسية، وحدود الإدخال متعدد الوسائط، وتفاصيل التسعير.
 
-### واجهة Seedance بدون وجوه (faces-off)
+### موافقة Seedance
 
-نماذج Seedance العامة **لا** تدعم تجهيز أصول الوجه ولا إثبات موافقة الوجه. قد تُرفض المدخلات التي تحتوي على وجوه في المنبع. لتوليد likeness / أصول الوجه استخدم تطبيق Venice أو Studio. راجع [دليل Seedance](/guides/media/seedance-2-0).
+عندما يحتوي طلب image-to-video أو reference-to-video في Seedance على وجه بشري، تُعيد الـ API استجابة `409 needs_consent` مع نص السياسة، وتعيد إرسال الطلب نفسه مع شهادة `consents.seedance`. راجع [دليل موافقة Seedance لوسائط الوجوه](/guides/media/seedance-face-consent) للتدفق الكامل وكائن الموافقة وإزالة التكرار والإلغاء.
 
 ### تكبير الفيديو
 
diff --git a/ar/guides/integrations/venice-video-harness.mdx b/ar/guides/integrations/venice-video-harness.mdx
index eec52cf4..44e2a326 100644
--- a/ar/guides/integrations/venice-video-harness.mdx
+++ b/ar/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ curl -L -o ~/.cache/whisper.cpp/ggml-base.en.bin \
 
 | الدور | النموذج الافتراضي | متى يُستخدم |
 |---|---|---|
-| لقطات الشخصيات (1-2 شخصية) | `seedance-2-0-reference-to-video-basic` | R2V افتراضي مع `reference_image_urls` مسطّحة، ووسوم `@Image`، حتى 15 ثانية، وصوت ستيريو أصلي |
+| لقطات الشخصيات (1-2 شخصية) | `seedance-2-0-reference-to-video` | R2V افتراضي مع `reference_image_urls` مسطّحة، ووسوم `@Image`، حتى 15 ثانية، وصوت ستيريو أصلي |
 | لقطات الشخصيات (3+ شخصيات) | `kling-o3-standard-reference-to-video` | بديل احتياطي تلقائي مع `elements` مهيكلة لهوية متعدّدة الشخصيات |
-| تأسيسي / مزاجي / حركة | `seedance-2-0-image-to-video-basic` | بدون شخصيات؛ جودة سينمائية ملحمية، حتى 15 ثانية |
+| تأسيسي / مزاجي / حركة | `seedance-2-0-image-to-video` | بدون شخصيات؛ جودة سينمائية ملحمية، حتى 15 ثانية |
 
 يمكن تجاوز هذه القيم لكل مشروع عبر `series.json → videoDefaults`. لاستهداف عائلة غير Seedance (مثل الحسابات التي تفتقر إلى وصول Seedance)، اضبط `videoDefaults` إلى `kling-o3-standard-reference-to-video` و`veo3.1-fast-image-to-video`.
 
diff --git a/ar/guides/media/seedance-2-0.mdx b/ar/guides/media/seedance-2-0.mdx
index a94123a0..9c7eacf0 100644
--- a/ar/guides/media/seedance-2-0.mdx
+++ b/ar/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 و 2.5"
-description: "أنشئ وحرّر ومدّد والصِق مقاطع الفيديو مع Seedance 2.0 و 2.5 على Venice - تدفقات النص/الصورة/المرجع، وسلوك واجهة faces-off، وحدود الوسائط المتعددة حسب العائلة."
-'og:title': "دليل Seedance 2.0 و 2.5 | وثائق Venice API"
-'og:description': "ولِّد وحرِّر ومدِّد وادمج مقاطع الفيديو باستخدام Seedance 2.0 و 2.5 عبر Venice API. أنماط سير العمل وfaces-off والحدود والتسعير والأمثلة."
+title: "Seedance 2.0"
+description: "أنشئ، حرّر، مدّد، ولصِّق مقاطع الفيديو مع Seedance 2.0 على Venice — تدفقات نص وصورة ومرجع إلى فيديو على endpoint واحد، مع أمثلة مشروحة."
+'og:title': "دليل Seedance 2.0 | وثائق Venice API"
+'og:description': "ولِّد وحرِّر ومدِّد وادمج مقاطع الفيديو باستخدام Seedance 2.0 عبر Venice API. أنماط سير العمل وحدود الإدخال متعدد الوسائط والتسعير وأمثلة عملية."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0 هو نموذج فيديو رائد متعدد الوسائط، متاح على Venice كعائلة من ثلاثة متغيّرات لتوليد الفيديو من النص والصورة والمراجع. متغيّر **reference-to-video** قويّ بشكل غير معتاد: نقطة نهاية واحدة ومعرّف نموذج واحد يتعاملان مع **أربعة سير عمل مميّزة** (Reference، Edit، Extend، Stitch) — ويُستنتج سير العمل من **شكل التعليمة**.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+يستعرض هذا الدليل المتغيّرات، وسير العمل الأربعة مع التعليمات المعيارية، وحدود الإدخال متعدد الوسائط، والتسعير، وأمثلة `curl` كاملة.
 
-<Warning>
-**الوجوه غير مدعومة على واجهة Seedance العامة.** لا تشغّل Venice تجهيز أصول الوجه ولا إثبات موافقة الوجه (`consents.seedance` / `needs_consent`) لنماذج Seedance العامة. قد تُرفض المدخلات التي تحتوي على وجوه في المنبع كخطأ سياسة محتوى / مزوّد. لتوليد likeness / أصول الوجه استخدم تطبيق Venice أو Studio.
-</Warning>
-
-## Variants
+## المتغيّرات
 
-| Model ID | Variant | Output resolutions | Notes |
+| معرّف النموذج | المتغيّر | دقّات الإخراج | ملاحظات |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | تعليمة نصية فقط |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | تثبيت بصورة الإطار الأول (واختياريًا الإطار الأخير) |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | حتى 9 صور مرجعية + 3 فيديوهات مرجعية + 3 مانحات صوتية. يدعم Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | فئة أسرع بجودة أقل |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | فئة أسرع بجودة أقل |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | فئة أسرع بجودة أقل؛ نفس مجموعة سير العمل |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+جميع المتغيّرات غير متزامنة. أرسل عبر `POST /api/v1/video/queue`، ثم استعلم بـ `POST /api/v1/video/retrieve` حتى يصبح جسم الاستجابة `video/mp4`. راجع [توليد الفيديو](/guides/media/video-generation) لتدفق الطابور العام.
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## نموذج «نموذج واحد، أربعة سير عمل»
 
-## The "one model, four workflows" model
+متغيّر reference-to-video (`seedance-2-0-reference-to-video` وشقيقه Fast) هو النموذج الأساسي ذاته يخدم أربع مهام مختلفة. **يستنتج النموذج المهمة من بادئة التعليمة وشكل مدخلاتك.** لا يوجد حقل `task` أو `workflow` — فبنية التعليمة هي ما يحدّد التوجيه.
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| سير العمل | ما يفعله | بادئة التعليمة | المدخلات |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | توليد فيديو جديد باستخدام الملفات المرجعية المرفوعة كمانحات للموضوع / الحركة / الأسلوب / الصوت | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | نص + ≥1 مرجع صورة أو فيديو (0–9 صور، 0–3 فيديوهات)، بالإضافة اختياريًا إلى ما يصل إلى 3 مانحات صوتية |
+| **Edit** | تعديل فيديو إدخال واحد مع الحفاظ على الباقي | `Strictly edit <Video 1>, changing its ...` | فيديو إدخال واحد + نص (صور اختيارية للتثبيت) |
+| **Extend** | تمديد أمامي / خلفي لمقطع واحد | `Extend <Video 1>, generate ...` | فيديو إدخال واحد + نص |
+| **Stitch** | دمج 2–3 مقاطع مع انتقالات مُولَّدة تلقائيًا | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2–3 فيديوهات إدخال + نص |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+**بنية التعليمة معيارية وحسّاسة لحالة الأحرف**: أقواس زاوية، حرف أول كبير، مسافة واحدة قبل الرقم — `<Video 1>`، `<Image 1>`، `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## أنماط سير العمل
 
-### Reference workflow
+### سير العمل Reference
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+استخدم الملفات المرجعية المرفوعة كـ**مانحات** — للموضوع، والمشهد، والحركة، والأسلوب، ونبرة الصوت — لتوليد فيديو جديد كليًا.
 
-**Canonical prompt patterns**:
+**أنماط التعليمة المعيارية**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,23 +51,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**أمثلة**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (يجب اقتران المانحات الصوتية بمرجع صورة أو فيديو واحد على الأقل — الصوت وحده يُرفض)
 
-### Edit workflow
+### سير العمل Edit
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+عدّل فيديو إدخال واحد. **كل ما لا يُذكر صراحة في التعليمة يُحفظ.** استخدم هذا عندما تريد تغييرًا محليًا (استبدال موضوع، تغيير طقس/لون، إضافة/إزالة عنصر) بدلًا من فيديو جديد كليًا.
 
-**Canonical prompt pattern**:
+**نمط التعليمة المعياري**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**أنماط فرعية لتحكّم أدق**:
 
 ```
 Add Elements:
@@ -89,57 +80,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**أمثلة**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+يجمع المثال الأخير بين Edit ومرجع صورة — وهذا مسموح تمامًا، إذ يستخدم النموذج `<Image 1>` كمانح بصري للاستبدال.
 
-### Extend workflow
+### سير العمل Extend
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+استمر بمقطع واحد للأمام أو للخلف زمنيًا. **افتراضيًا يعيد Seedance المحتوى الجديد فقط** — وليس الإدخال الأصلي متصلًا بالامتداد. هذا مقصود لاستمرارية الانتقال؛ وإذا أردت الحفاظ على المقطع الأصلي إلى جانب الامتداد، فقل ذلك صراحة:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← يحفظ الإدخال في البداية
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← يحفظ الإدخال في النهاية
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+التعامل مع الانتقال: يستخرج النموذج تلقائيًا إطارات الانتقال للمزج السلس، ولا تُعاد توليد الأجزاء الأصلية من فيديو الإدخال.
 
-**Examples**:
+**أمثلة**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### سير العمل Stitch (إكمال المسار)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+اربط 2–3 مقاطع إدخال بانتقالات مُولَّدة بالذكاء الاصطناعي. يجب ألّا تتجاوز المدة الإجمالية المجمّعة للإدخال 15 ثانية.
 
-**Canonical prompt pattern**:
+**نمط التعليمة المعياري**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**أمثلة**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+يقصّ النموذج تلقائيًا المقاطع الموصِلة عند نقاط الالتقاء لضمان الاستمرارية.
 
 ---
 
-## Universal prompt formula
+## صيغة التعليمة الموحّدة
 
-Across all four workflows, the recommended authoring formula is:
+عبر سير العمل الأربعة، الصيغة الموصى بها للتأليف هي:
 
 ```
 Subject + Motion + Environment (Optional)
@@ -148,77 +139,79 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subject + Motion**: الأساس المنطقي — حدّد «مَن» يؤدّي «أي فعل»
+- **Environment + Aesthetics**: الخلفية المكانية، الإضاءة، الأسلوب البصري
+- **Camera**: نوع لقطة أو حركة صريحة
+- **Audio**: مؤثرات صوتية محيطة أو توجيه صوتي لإخراج غامر
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+وضع هذه فوق بادئة سير العمل (مثل `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) يُنتج أعلى جودة من المخرجات.
 
 ---
 
-## Multimodal input limits
+## حدود الإدخال متعدد الوسائط
+
+القيم أدناه هي ما يقبله Venice API. الطلبات خارج هذه النطاقات تُرفض على طبقة المخطط برمز 400 قبل الوصول إلى الاستدلال.
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### الصور
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| القيد | القيمة |
+|---|---|
+| طرق الإدخال | URL (`http://`، `https://`) أو رابط بيانات Base64 (`data:image/...`) |
+| الصيغ | `.jpeg`، `.png`، `.webp`، `.bmp`، `.tiff`، `.gif`، `.heic`، `.heif` |
+| نسبة العرض إلى الارتفاع (W / H) | حصرية `(0.4, 2.5)` |
+| أدنى ضلع | ≥ 300 بكسل |
+| عدد الصور: I2V الإطار الأول | 1 |
+| عدد الصور: I2V الإطار الأول + الأخير | 2 |
+| **عدد الصور: R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### الفيديوهات
 
-| Constraint | Value |
+| القيد | القيمة |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| طرق الإدخال | URL (`http://`، `https://`) أو رابط بيانات Base64 (`data:video/...`) |
+| الصيغ | `.mp4`، `.mov` |
+| ترميزات الفيديو | H.264 / AVC، H.265 / HEVC |
+| ترميزات الصوت (في الحاوية) | AAC، MP3 |
+| **مدة كل مقطع** | `[2, 15]` ث (شامل) |
+| **الحد الأقصى لعدد المقاطع** | 3 (R2V / Stitch / Extend) |
+| **إجمالي المدة المجمّعة** | ≤ 15 ث عبر جميع المقاطع |
+| **حجم كل مقطع** | ≤ 50 ميغابايت |
+
+### الصوت
+
+| القيد | القيمة |
+|---|---|
+| طرق الإدخال | URL (`http://`، `https://`) أو رابط بيانات Base64 (`data:audio/...`) |
+| الصيغ | `.wav`، `.mp3` |
+| **مدة كل مقطع** | `[2, 15]` ث |
+| **الحد الأقصى لعدد المقاطع** | 3 |
+| **إجمالي المدة المجمّعة** | ≤ 15 ث عبر جميع المقاطع |
+| **حجم كل مقطع** | ≤ 15 ميغابايت |
+
+الصوت المرجعي مدعوم على متغيّرات R2V فقط. تُمرَّر كل مدخلة إلى النموذج كعنصر محتوى بدور `role: "reference_audio"` تشير إليه التعليمة بـ `<Audio 1>`، `<Audio 2>`، `<Audio 3>` — ويستخدم النموذج كل مقطع لنبرة صوت أو مؤثرات صوتية أو موسيقى خلفية اعتمادًا على كيفية صياغة التعليمة. حقل `audio_url` المفرد القديم يُربط بنفس شكل المحتوى وهو الآن مكافئ لتمرير `reference_audio_urls` بعنصر واحد.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**لا يمكن أن يكون `reference_audio_urls` المدخل المرجعي الوحيد.** يتطلب النموذج مرجع صورة أو فيديو واحدًا على الأقل إلى جانب أي مانح صوتي. اقرن `reference_audio_urls` بـ `reference_image_urls` أو `reference_video_urls` أو `image_url` أو `video_url` — تُرفض الطلبات الصوتية فقط.
 </Warning>
 
-### Request size
+### حجم الطلب
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+تقبل نقطة نهاية الطابور أجسام JSON حتى **35 ميغابايت**. روابط البيانات المضمّنة لفيديوهات كبيرة قد تتجاوز هذا — لـ Stitch متعدد المقاطع خصوصًا، فضّل الـ URLs على base64 المضمّن.
 
 ---
 
-## Pricing
+## التسعير
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+استدعِ `POST /api/v1/video/quote` للحصول على عرض سعر لشكل طلب معيّن قبل إرساله إلى `/video/queue`. نقطة نهاية الاقتباس هي المصدر المعتمد الوحيد؛ قد تتغيّر تفاصيل التسعير ولا ينبغي تخزينها مؤقتًا أو تكرارها على جانب العميل.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+عند تضمين فيديو(هات) مرجعية في الطلب، مرّر أيضًا `reference_video_total_duration` (مجموع كل مدد المقاطع المرجعية بالثواني) لتطابُق عرض السعر مع ما سيحاسب عليه `/video/queue`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,18 +221,18 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## أمثلة كاملة
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+جميع الأمثلة تفترض ضبط `VENICE_API_KEY` في البيئة.
 
-### Text-to-video
+### من نص إلى فيديو
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### من صورة إلى فيديو (الإطار الأول)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+لا يقبل `seedance-2-0-image-to-video` (ومتغيّره Fast) حقل `aspect_ratio` — إذ تُشتق نسبة الإخراج تلقائيًا من أبعاد صورة الإدخال. تمرير الحقل يُعيد 400 مع رسالة *"This model does not support aspect_ratio"*. استخدم متغيّرات T2V أو R2V إذا كنت تحتاج تحكّمًا صريحًا في نسبة العرض إلى الارتفاع.
 </Note>
 
-### Reference workflow — subject donor
+### سير عمل Reference — مانح موضوع
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### سير عمل Reference — مانح موضوع + صوت
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### سير عمل Edit
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### سير عمل Edit مع تثبيت بصورة
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### Extend إلى الأمام
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -400,14 +344,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clips)
+### Stitch (3 مقاطع)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### الاستعلام عن الاكتمال
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+بعد كل إرسال إلى الطابور، احفظ `queue_id` المُعاد واستعلم بـ `/video/retrieve` حتى يصبح جسم الاستجابة `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+تكون الاستجابة JSON (`{ "status": "queued" | "running" | "failed", ... }`) حتى تكتمل المهمة، عندها يتحوّل جسم الاستجابة إلى بايتات `video/mp4`. راجع [توليد الفيديو](/guides/media/video-generation) لنمط الاستعلام الكامل.
 
 ---
 
-## Troubleshooting
+## استكشاف الأخطاء وإصلاحها
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+يجب أن تتضمن طلبات reference-to-video واحدًا على الأقل من `reference_image_urls` أو `reference_video_urls` أو `image_references` أو `video_references`. التوليد النصي البحت ليس سير عمل R2V صالحًا — استخدم `seedance-2-0-text-to-video` بدلًا من ذلك. لا يكفي `reference_audio_urls` وحده (راجع قسم الصوت أعلاه).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+يحدّد النموذج الفيديوهات المرجعية بـ 3. إن احتجت إلى مقاطع أكثر، نفّذ Stitch أولًا (3 → 1)، ثم استخدم الإخراج كمرجع لمتابعة لاحقة.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / المجموع `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+مدة كل مقطع `[2, 15]` ثانية **شاملاً**؛ ومجموع كل الفيديوهات المرجعية محدّد أيضًا بـ 15 ثانية. اقتص المقاطع على جانب العميل قبل الإرسال.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### تعليمة توجِّه إلى سير العمل الخاطئ
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+يُستنتج سير العمل من بنية التعليمة. أخطاء توجيه شائعة:
 
-### Face / likeness inputs rejected
+- ترغب في **Extend** لكنك تكتب `Refer to ...` ← يعامل النموذج فيديوك كـ*مانح*، وليس كقماشة للاستمرار
+- ترغب في **Stitch** لكنك تكتب `Refer to ...` ← يختار النموذج أحدها كمانح، ويتجاهل الباقي
+- ترغب في **Edit** لكنك تكتب `Generate a video based on <Video 1>` ← غامض؛ قد يلجأ النموذج إلى Reference
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+استخدم البادئات المعيارية كما هي مكتوبة بالضبط: `Strictly edit <Video 1>, ...`، `Extend <Video 1>, ...`، `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### عرض السعر لا يطابق المبلغ المُجدوَل
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+إذا ضمّنت فيديو مرجعيًا ولم تمرّر `reference_video_total_duration` إلى `/video/quote`، فقد يختلف عرض السعر عن المبلغ المُجدوَل. مرّر دومًا `reference_video_total_duration` (مجموع كل مدد المقاطع المرجعية بالثواني) عند وجود فيديوهات مرجعية.
 
 ---
 
-## References
+## مراجع
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- نقطة نهاية طابور فيديو Venice: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- نقطة نهاية الاقتباس في Venice: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- دليل مرافق: [المرجع إلى الفيديو](/guides/media/reference-to-video) (يغطي Kling O3 + Grok Imagine R2V)
+- دليل مرافق: [توليد الفيديو](/guides/media/video-generation) (نظرة عامة على الطابور / الاستعلام)
diff --git a/ar/guides/media/video-generation.mdx b/ar/guides/media/video-generation.mdx
index 5d538e5a..a8f8afb0 100644
--- a/ar/guides/media/video-generation.mdx
+++ b/ar/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ await fetch(`${BASE_URL}/video/complete`, {
 عرض السعر بالدولار الأمريكي.
 
 <Info>
-**معاملات عرض السعر هي مدخلات تسعير فقط.** الحقول التي ترسلها إلى `/video/quote` — بما في ذلك `aspect_ratio` و`audio` و`reference_video_total_duration` — تُستخدم بحتًا لحساب السعر. **لا** تُمرَّر إلى `/video/queue` ولا تؤثر في الفيديو المولَّد. يُقبل إرسال `aspect_ratio` إلى quote حتى بالنسبة للنماذج (مثل `seedance-2-0-image-to-video-basic`) التي ترفض `aspect_ratio` وقت التوليد، لأن نماذج الصورة إلى فيديو تشتق نسبة العرض إلى الارتفاع للإخراج من صورة الإدخال.
+**معاملات عرض السعر هي مدخلات تسعير فقط.** الحقول التي ترسلها إلى `/video/quote` — بما في ذلك `aspect_ratio` و`audio` و`reference_video_total_duration` — تُستخدم بحتًا لحساب السعر. **لا** تُمرَّر إلى `/video/queue` ولا تؤثر في الفيديو المولَّد. يُقبل إرسال `aspect_ratio` إلى quote حتى بالنسبة للنماذج (مثل `seedance-2-0-image-to-video`) التي ترفض `aspect_ratio` وقت التوليد، لأن نماذج الصورة إلى فيديو تشتق نسبة العرض إلى الارتفاع للإخراج من صورة الإدخال.
 </Info>
 
 ### تسعير المرجع إلى فيديو
diff --git a/de/api-reference/endpoint/video/queue.mdx b/de/api-reference/endpoint/video/queue.mdx
index 3594795d..00033679 100644
--- a/de/api-reference/endpoint/video/queue.mdx
+++ b/de/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Rufen Sie `/video/quote` auf, um einen Preisvoranschlag zu erhalten, und fragen
 
 Private Modelle geben außerdem eine `download_url` für das fertige Video zurück. Es handelt sich um eine kurzlebige Delivery-URL (einige Wiederholungen sind in Ordnung, wenn ein Download abbricht); Details und das optionale `DELETE` aus Datenschutzgründen finden Sie im [Video Generation Guide](/guides/media/video-generation#private-download-links).
 
-### Seedance 2.0 & 2.5
+### Seedance 2.0
 
-Für `seedance-2-0-*-basic`- und `seedance-2-5-*-basic`-Modelle (Text-to-Video, Image-to-Video, Reference-to-Video sowie die Seedance-2.0-`-fast-*`-Varianten) siehe den [Seedance-2.0-&-2.5-Guide](/guides/media/seedance-2-0) zu den vier Workflows (Reference / Edit / Extend / Stitch), familienspezifischen Multimodal-Limits, Faces-off-API-Verhalten und Preisen.
+Für `seedance-2-0-*`-Modelle (text-to-video, image-to-video, reference-to-video plus die `-fast-*`-Varianten) siehe den [Seedance-2.0-Leitfaden](/guides/media/seedance-2-0) für das Vier-Workflow-Modell (Reference / Edit / Extend / Stitch), kanonische Prompt-Muster, multimodale Eingabegrenzen und Preisdetails.
 
-### Seedance Faces-off API
+### Seedance-Einwilligung
 
-Öffentliche Seedance-Modelle unterstützen **kein** Face-Asset-Provisioning und keine Face-Consent-Attestierung. Eingaben mit Gesichtern können upstream abgelehnt werden. Für Likeness-/Face-Asset-Generierung die Venice-App oder Studio nutzen. Details im [Seedance-Guide](/guides/media/seedance-2-0).
+Wenn eine Seedance-Image- oder Reference-to-Video-Anfrage ein menschliches Gesicht enthält, gibt die API ein `409 needs_consent` mit dem Richtlinientext zurück, und Sie reichen dieselbe Anfrage erneut mit einer `consents.seedance`-Bestätigung ein. Den vollständigen Ablauf, das Consent-Objekt, Dedup und Widerruf finden Sie im [Seedance-Face-Media-Consent-Leitfaden](/guides/media/seedance-face-consent).
 
 ### Video-Upscaling
 
diff --git a/de/guides/integrations/venice-video-harness.mdx b/de/guides/integrations/venice-video-harness.mdx
index 3e08f32a..146921f0 100644
--- a/de/guides/integrations/venice-video-harness.mdx
+++ b/de/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Die Standardeinstellungen sind bewusst meinungsstark, weil Konsistenz das Ziel i
 
 | Rolle | Standardmodell | Wann verwendet |
 |---|---|---|
-| Charakter-Shots (1–2 Charaktere) | `seedance-2-0-reference-to-video-basic` | Standard-R2V mit flachen `reference_image_urls`, `@Image`-Tags, bis zu 15 s, nativer Stereo-Sound |
+| Charakter-Shots (1–2 Charaktere) | `seedance-2-0-reference-to-video` | Standard-R2V mit flachen `reference_image_urls`, `@Image`-Tags, bis zu 15 s, nativer Stereo-Sound |
 | Charakter-Shots (3+ Charaktere) | `kling-o3-standard-reference-to-video` | Auto-Fallback mit strukturierten `elements` für Multi-Charakter-Identität |
-| Establishing / Stimmung / Action | `seedance-2-0-image-to-video-basic` | Keine Charaktere; epische Kino-Qualität, bis zu 15 s |
+| Establishing / Stimmung / Action | `seedance-2-0-image-to-video` | Keine Charaktere; epische Kino-Qualität, bis zu 15 s |
 
 Diese sind pro Projekt über `series.json → videoDefaults` überschreibbar. Für eine Nicht-Seedance-Familie (z. B. Accounts ohne Seedance-Zugriff) `videoDefaults` auf `kling-o3-standard-reference-to-video` und `veo3.1-fast-image-to-video` setzen.
 
diff --git a/de/guides/media/seedance-2-0.mdx b/de/guides/media/seedance-2-0.mdx
index f1ef7bdd..175df530 100644
--- a/de/guides/media/seedance-2-0.mdx
+++ b/de/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 & 2.5"
-description: "Videos mit Seedance 2.0 und 2.5 auf Venice erzeugen, bearbeiten, verlängern und zusammenfügen - Text-/Bild-/Referenz-Workflows, Faces-off-API und familienspezifische Multimodal-Limits."
-'og:title': "Seedance 2.0 & 2.5 Guide | Venice API Docs"
-'og:description': "Videos mit Seedance 2.0 und 2.5 über die Venice API erzeugen. Workflows, Faces-off, Multimodal-Limits, Preise und Beispiele."
+title: "Seedance 2.0"
+description: "Generieren, bearbeiten, verlängern und stitchen Sie Videos mit Seedance 2.0 auf Venice — Text-, Image- und Reference-to-Video an einem Endpoint."
+'og:title': "Seedance 2.0-Guide | Venice API Docs"
+'og:description': "Videos mit Seedance 2.0 über die Venice-API generieren, editieren, verlängern und stitchen. Workflow-Muster, multimodale Limits, Preise und ausgearbeitete Beispiele."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0 ist ein Flaggschiff-Multimodal-Videomodell, das in Venice als Familie aus drei Varianten für Text-, Image- und Referenz-getriebene Videogenerierung verfügbar ist. Die **Reference-to-Video**-Variante ist außergewöhnlich mächtig: ein einziger Endpoint und eine einzige Modell-ID handhaben **vier verschiedene Workflows** (Reference, Edit, Extend, Stitch) — der Workflow wird aus der **Form deines Prompts** abgeleitet.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+Dieser Guide führt durch die Varianten, die vier Workflows mit kanonischen Prompts, die Multimodal-Input-Limits, Preise und vollständige `curl`-Beispiele.
 
-<Warning>
-**Gesichter werden auf der öffentlichen Seedance-API nicht unterstützt.** Venice führt für öffentliche Seedance-Modelle kein Face-Asset-Provisioning und keine Face-Consent-Attestierung (`consents.seedance` / `needs_consent`) aus. Eingaben mit Gesichtern können upstream als Content-Policy-/Provider-Fehler abgelehnt werden. Für Likeness-/Face-Asset-Generierung die Venice-App oder Studio nutzen.
-</Warning>
-
-## Variants
+## Varianten
 
-| Model ID | Variant | Output resolutions | Notes |
+| Modell-ID | Variante | Output-Auflösungen | Hinweise |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Nur Text-Prompt |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Erst- (und optional Letzt-)Frame-Bildgrundlage |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Bis zu 9 Referenzbilder + 3 Referenzvideos + 3 Referenz-Audio-Donor. Treibt Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Schneller, niedrigerer Fidelity-Tier |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Schneller, niedrigerer Fidelity-Tier |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Schneller, niedrigerer Fidelity-Tier; selber Workflow-Satz |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+Alle Varianten sind async. Submit per `POST /api/v1/video/queue`, dann `POST /api/v1/video/retrieve` pollen, bis der Response-Body `video/mp4` ist. Siehe [Videogenerierung](/guides/media/video-generation) für den allgemeinen Queue-Flow.
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## Das „Ein Modell, vier Workflows"-Modell
 
-## The "one model, four workflows" model
+Die Reference-to-Video-Variante (`seedance-2-0-reference-to-video` und das Fast-Gegenstück) ist dasselbe zugrundeliegende Modell, das vier verschiedene Aufgaben bedient. **Das Modell leitet die Aufgabe aus dem Prompt-Präfix und der Form deiner Inputs ab.** Es gibt kein `task`- oder `workflow`-Feld – die Prompt-Syntax ist das Routing.
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| Workflow | Was es tut | Prompt-Präfix | Inputs |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | Ein neues Video generieren, mit hochgeladenen Referenzdateien als Donor für Subjekt / Bewegung / Stil / Audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 Bild- ODER Video-Referenz (0–9 Bilder, 0–3 Videos), plus optional bis zu 3 Audio-Donor |
+| **Edit** | Ein einzelnes Eingabevideo modifizieren, den Rest erhalten | `Strictly edit <Video 1>, changing its ...` | 1 Eingabevideo + Text (Bilder optional als Grounding) |
+| **Extend** | Vorwärts-/Rückwärts-Verlängerung eines Clips | `Extend <Video 1>, generate ...` | 1 Eingabevideo + Text |
+| **Stitch** | 2–3 Clips mit automatisch generierten Übergängen verbinden | `<Video 1> + <Transition-Beschreibung> + followed by <Video 2> + ...` | 2–3 Eingabevideos + Text |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+Die **Prompt-Syntax ist kanonisch und case-sensitiv**: spitze Klammern, Großbuchstabe am Anfang, ein Leerzeichen vor der Zahl – `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## Workflow-Muster
 
-### Reference workflow
+### Reference-Workflow
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+Nutze die hochgeladenen Referenzdateien als **Donor** – Subjekt, Szene, Bewegung, Stil, Stimmtimbre –, um ein völlig neues Video zu generieren.
 
-**Canonical prompt patterns**:
+**Kanonische Prompt-Muster**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,165 +51,167 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**Beispiele**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (Audio-Donor müssen mit mindestens einer Bild- oder Video-Referenz gepaart sein – nur Audio wird abgelehnt)
 
-### Edit workflow
+### Edit-Workflow
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+Modifiziere ein einzelnes Eingabevideo. **Alles, was nicht explizit im Prompt genannt ist, bleibt erhalten.** Nutze dies, wenn du eine lokale Änderung möchtest (Subjekttausch, Wetter-/Farbänderung, Element-Add/Remove) statt eines komplett neuen Videos.
 
-**Canonical prompt pattern**:
+**Kanonisches Prompt-Muster**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**Sub-Muster für feinere Kontrolle**:
 
 ```
-Add Elements:
+Elemente hinzufügen:
   At [timestamp / timing] and [spatial location] of <Video 1>, add [description of intended element].
 
-Remove Elements:
+Elemente entfernen:
   Remove [element to be deleted] from <Video 1>, keeping the rest of the video content unchanged.
 
-Modify Elements:
+Elemente ändern:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**Beispiele**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+Das letzte Beispiel kombiniert Edit mit einer Bildreferenz – völlig legitim, das Modell nutzt `<Image 1>` als visuellen Donor für den Austausch.
 
-### Extend workflow
+### Extend-Workflow
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+Setze einen einzelnen Clip zeitlich vorwärts oder rückwärts fort. **Standardmäßig liefert Seedance nur den neuen Inhalt** – nicht das Original-Input verkettet mit der Verlängerung. Das ist für die Übergangskontinuität gewollt; soll der Eingabeclip neben der Verlängerung erhalten bleiben, sage es explizit:
 
 ```
-Extend <Video 1>, generate [description of extended content]
-Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, generate [Beschreibung des Verlängerungsinhalts]
+Extend <Video 1> backward, [Beschreibung des Verlängerungsinhalts]
+Extend <Video 1>, start with <Video 1>, then [Beschreibung des Verlängerungsinhalts]      ← Input am Anfang erhalten
+Extend <Video 1> backward, [Beschreibung], and then end with <Video 1>                    ← Input am Ende erhalten
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+Übergangsbehandlung: Das Modell extrahiert automatisch die Übergangsframes für nahtloses Blending; die Originalsegmente des Eingabevideos werden nicht neu generiert.
 
-**Examples**:
+**Beispiele**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Stitch-Workflow (Track-Completion)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+Verbinde 2–3 Eingangsclips mit KI-generierten Übergängen. Die kombinierte Eingangsdauer muss ≤ 15 s sein.
 
-**Canonical prompt pattern**:
+**Kanonisches Prompt-Muster**:
 
 ```
-<Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
+<Video 1> + [Transition-Beschreibung] + followed by <Video 2> [+ [Transition-Beschreibung] + followed by <Video 3>]
 ```
 
-**Examples**:
+**Beispiele**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+Das Modell trimmt verbindende Segmente an den Übergängen automatisch für Kontinuität.
 
 ---
 
-## Universal prompt formula
+## Universelle Prompt-Formel
 
-Across all four workflows, the recommended authoring formula is:
+Über alle vier Workflows hinweg lautet die empfohlene Schreibformel:
 
 ```
-Subject + Motion + Environment (Optional)
-       + Camera Movement / Cut (Optional)
-       + Aesthetic Description (Optional)
-       + Audio (Optional)
+Subjekt + Bewegung + Umgebung (optional)
+       + Kamerabewegung / Schnitt (optional)
+       + Ästhetik-Beschreibung (optional)
+       + Audio (optional)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subjekt + Bewegung**: das logische Fundament – definiere „Wer" führt „Welche Aktion" aus
+- **Umgebung + Ästhetik**: räumlicher Hintergrund, Beleuchtung, Bildstil
+- **Kamera**: expliziter Shot-Typ oder Bewegung
+- **Audio**: Ambient-Soundeffekte oder Stimmrichtung für immersiven Output
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+Wenn du das auf ein Workflow-Präfix legst (z. B. `Strictly edit <Video 1>, changing its <Subjekt + Bewegung + Umgebung + ...>`), bekommst du die qualitativ besten Ergebnisse.
 
 ---
 
-## Multimodal input limits
+## Multimodale Input-Limits
+
+Die unten genannten Werte sind das, was die Venice-API akzeptiert. Anfragen außerhalb dieser Bereiche werden auf der Schema-Ebene mit 400 abgelehnt, bevor sie die Inferenz erreichen.
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### Bilder
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| Constraint | Wert |
+|---|---|
+| Input-Methoden | URL (`http://`, `https://`) oder Base64-Data-URL (`data:image/...`) |
+| Formate | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Aspect Ratio (B / H) | exklusiv `(0.4, 2.5)` |
+| Mindestseite | ≥ 300 px |
+| Bildanzahl: I2V Erst-Frame | 1 |
+| Bildanzahl: I2V Erst- + Letzt-Frame | 2 |
+| **Bildanzahl: R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### Videos
 
-| Constraint | Value |
+| Constraint | Wert |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Input-Methoden | URL (`http://`, `https://`) oder Base64-Data-URL (`data:video/...`) |
+| Formate | `.mp4`, `.mov` |
+| Video-Codecs | H.264 / AVC, H.265 / HEVC |
+| Audio-Codecs (im Container) | AAC, MP3 |
+| **Dauer pro Clip** | `[2, 15]` s (inklusiv) |
+| **Max. Clip-Anzahl** | 3 (R2V / Stitch / Extend) |
+| **Gesamtdauer kombiniert** | ≤ 15 s über alle Clips |
+| **Größe pro Clip** | ≤ 50 MB |
+
+### Audio
+
+| Constraint | Wert |
+|---|---|
+| Input-Methoden | URL (`http://`, `https://`) oder Base64-Data-URL (`data:audio/...`) |
+| Formate | `.wav`, `.mp3` |
+| **Dauer pro Clip** | `[2, 15]` s |
+| **Max. Clip-Anzahl** | 3 |
+| **Gesamtdauer kombiniert** | ≤ 15 s über alle Clips |
+| **Größe pro Clip** | ≤ 15 MB |
+
+Referenz-Audio wird nur in den R2V-Varianten unterstützt. Jeder Eintrag wird an das Modell als Content-Item mit `role: "reference_audio"` weitergegeben, das der Prompt als `<Audio 1>`, `<Audio 2>`, `<Audio 3>` adressiert – das Modell nutzt jeden Clip je nach Prompt-Rahmen für Stimmtimbre, Soundeffekte oder Hintergrundmusik. Das veraltete Singular-Feld `audio_url` mappt auf dieselbe Content-Form und ist nun äquivalent zur Übergabe eines einzelelementigen `reference_audio_urls`.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls` kann nicht der einzige Referenz-Input sein.** Das Modell verlangt mindestens eine Bild- oder Videoreferenz neben jedem Audio-Donor. Paare `reference_audio_urls` mit `reference_image_urls`, `reference_video_urls`, `image_url` oder `video_url` – reine Audio-Submissions werden abgelehnt.
 </Warning>
 
-### Request size
+### Request-Größe
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+Der Queue-Endpoint akzeptiert JSON-Bodies bis zu **35 MB**. Inline-Data-URLs für große Videos können das überschreiten – besonders bei Multi-Clip-Stitch URLs gegenüber Inline-Base64 bevorzugen.
 
 ---
 
-## Pricing
+## Preise
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+Rufe `POST /api/v1/video/quote` auf, um ein Quote für eine bestimmte Request-Form zu erhalten, bevor du es an `/video/queue` schickst. Der Quote-Endpoint ist die einzige maßgebliche Quelle; Pricing-Details können sich ändern und sollten nicht client-seitig gecached oder dupliziert werden.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+Wenn Referenzvideos Teil des Requests sind, übergib auch `reference_video_total_duration` (die Summe aller Referenzclip-Dauern in Sekunden), damit das Quote dem entspricht, was `/video/queue` abrechnet:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,18 +221,18 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## Vollständige Beispiele
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+Alle Beispiele setzen voraus, dass `VENICE_API_KEY` in der Umgebung gesetzt ist.
 
-### Text-to-video
+### Text-to-Video
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### Image-to-Video (Erst-Frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video` (und seine Fast-Variante) **akzeptieren `aspect_ratio` nicht** – das Output-Aspect-Ratio wird automatisch aus den Dimensionen des Eingabebilds abgeleitet. Wird das Feld übergeben, kommt 400 mit *„This model does not support aspect_ratio"*. Verwende die T2V- oder R2V-Varianten, wenn du explizite Aspect-Ratio-Kontrolle brauchst.
 </Note>
 
-### Reference workflow — subject donor
+### Reference-Workflow — Subjekt-Donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Reference-Workflow — Subjekt + Audio-Donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Edit-Workflow
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### Edit-Workflow mit Bild-Grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### Extend (vorwärts)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -400,14 +344,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clips)
+### Stitch (3 Clips)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### Polling auf Fertigstellung
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+Speichere nach jedem Queue-Submit die zurückgegebene `queue_id` und pollte `/video/retrieve`, bis der Response-Body `video/mp4` ist:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+Die Antwort ist JSON (`{ "status": "queued" | "running" | "failed", ... }`), bis der Job abgeschlossen ist; dann wechselt der Response-Body auf `video/mp4`-Bytes. Siehe [Videogenerierung](/guides/media/video-generation) für das vollständige Polling-Muster.
 
 ---
 
-## Troubleshooting
+## Fehlerbehebung
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Reference-to-Video-Submissions müssen mindestens eines von `reference_image_urls`, `reference_video_urls`, `image_references` oder `video_references` enthalten. Reine Text-Generierung ist kein gültiger R2V-Workflow – verwende stattdessen `seedance-2-0-text-to-video`. `reference_audio_urls` allein reicht nicht (siehe Audio-Abschnitt oben).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+Das Modell deckelt Referenzvideos bei 3. Wenn du mehr Clips brauchst, mach zuerst einen Stitch (3 → 1) und nutze den Output als Referenz für einen Folge-Job.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / Aggregat `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+Die Dauer pro Clip ist `[2, 15]` Sekunden **inklusiv**; die Summe über alle Referenzvideos ist ebenfalls auf 15 Sekunden gedeckelt. Clips vor dem Submit client-seitig trimmen.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### Prompt routet auf den falschen Workflow
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Der Workflow wird aus der Prompt-Syntax abgeleitet. Häufige Fehlrouten:
 
-### Face / likeness inputs rejected
+- **Extend** gemeint, aber `Refer to ...` geschrieben → Modell behandelt dein Video als *Donor*, nicht als fortzusetzende Canvas
+- **Stitch** gemeint, aber `Refer to ...` geschrieben → Modell wählt einen als Donor, ignoriert die anderen
+- **Edit** gemeint, aber `Generate a video based on <Video 1>` geschrieben → mehrdeutig; Modell fällt evtl. auf Reference zurück
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+Nutze die kanonischen Präfixe exakt wie geschrieben: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### Quote stimmt nicht mit dem Queue-Betrag überein
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+Wenn du ein Referenzvideo eingebunden, aber `reference_video_total_duration` nicht an `/video/quote` übergeben hast, können Quote und Queue-Betrag abweichen. Bei vorhandenen Referenzvideos immer `reference_video_total_duration` (Summe aller Referenzclip-Dauern in Sekunden) übergeben.
 
 ---
 
-## References
+## Referenzen
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Venice Video-Queue-Endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice Quote-Endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Begleitender Guide: [Reference to Video](/guides/media/reference-to-video) (deckt Kling O3 + Grok Imagine R2V ab)
+- Begleitender Guide: [Videogenerierung](/guides/media/video-generation) (Queue-/Polling-Übersicht)
diff --git a/de/guides/media/video-generation.mdx b/de/guides/media/video-generation.mdx
index f494cb4b..ec9bdd18 100644
--- a/de/guides/media/video-generation.mdx
+++ b/de/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Exakte Kosten vor der Generierung. Nur die Pricing-Inputs senden (`model`, `dura
 Das Quote ist in USD.
 
 <Info>
-**Quote-Parameter sind ausschließlich Preis-Inputs.** Felder, die Sie an `/video/quote` senden – darunter `aspect_ratio`, `audio` und `reference_video_total_duration` – dienen ausschließlich der Preisberechnung. Sie werden **nicht** an `/video/queue` weitergeleitet und haben keinen Einfluss auf das generierte Video. `aspect_ratio` an Quote zu senden ist auch bei Modellen zulässig (z. B. `seedance-2-0-image-to-video-basic`), die `aspect_ratio` bei der Generierung ablehnen, da Image-to-Video-Modelle das Ausgabe-Seitenverhältnis aus dem Eingabebild ableiten.
+**Quote-Parameter sind ausschließlich Preis-Inputs.** Felder, die Sie an `/video/quote` senden – darunter `aspect_ratio`, `audio` und `reference_video_total_duration` – dienen ausschließlich der Preisberechnung. Sie werden **nicht** an `/video/queue` weitergeleitet und haben keinen Einfluss auf das generierte Video. `aspect_ratio` an Quote zu senden ist auch bei Modellen zulässig (z. B. `seedance-2-0-image-to-video`), die `aspect_ratio` bei der Generierung ablehnen, da Image-to-Video-Modelle das Ausgabe-Seitenverhältnis aus dem Eingabebild ableiten.
 </Info>
 
 ### Preisberechnung bei Reference-to-Video
diff --git a/docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md b/docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md
deleted file mode 100644
index 1c7d99c6..00000000
--- a/docs/superpowers/plans/2026-08-08-api-103-seedance-faces-off-docs.md
+++ /dev/null
@@ -1,25 +0,0 @@
-# API-103 Seedance faces-off docs Implementation Plan
-
-> **For agentic workers:** Implement task-by-task. Steps use checkbox syntax.
-
-**Goal:** Re-enable Seedance public docs for faces-off API (2.0 + 2.5 limits), leave face-consent off-nav unchanged, fix queue reference blurbs.
-
-**Architecture:** Mintlify MDX guides + `docs.json` nav; content sourced from outerface Seedance model defs / ENG-950 product rules.
-
-**Tech Stack:** Mintlify, MDX, JSON nav
-
-## Global Constraints
-
-- Keep URL `guides/media/seedance-2-0`; retitle for 2.0 + 2.5
-- Face-consent pages: no content edits; stay off-nav
-- Do not edit `swagger.yaml`
-- Split 2.0 vs 2.5 multimodal limits; never present 2.0 caps as applying to 2.5
-- Full i18n for guide + queue blurb + nav
-
-## Tasks
-
-- [ ] Task 1: Update English `guides/media/seedance-2-0.mdx`
-- [ ] Task 2: Update English `api-reference/endpoint/video/queue.mdx`
-- [ ] Task 3: Re-enable Seedance guide in all `docs.json` locale navs
-- [ ] Task 4: Port guide + queue changes to ar/de/es/fr/it/ko/pt-BR/zh
-- [ ] Task 5: Grep for remaining live consent how-tos; sanity-check nav
diff --git a/docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md b/docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md
deleted file mode 100644
index d226f2cc..00000000
--- a/docs/superpowers/specs/2026-08-08-api-103-seedance-faces-off-docs-design.md
+++ /dev/null
@@ -1,79 +0,0 @@
-# API-103 — Seedance faces-off public docs (re-enable)
-
-**Date:** 2026-08-08  
-**Status:** Approved  
-**Repo:** `veniceai/api-docs`  
-**Linear:** [API-103](https://linear.app/venice/issue/API-103)  
-**Depends on:** [ENG-950](https://linear.app/venice/issue/ENG-950) (hard-cut + listed no-face API twins)
-
-## Goal
-
-Public Venice API docs match faces-off Seedance API behavior after ENG-950, document Seedance **2.0 and 2.5** (with correct multimodal limits), and re-enable the Seedance guide in the public nav.
-
-## Decisions
-
-| Topic | Decision |
-| --- | --- |
-| Face-consent page | Keep file + locale copies; **leave content unchanged**; **stay off-nav** |
-| Seedance guide | Retitle for 2.0 + 2.5; keep URL `guides/media/seedance-2-0` |
-| Nav re-enable | On this docs PR merge (do not wait for a second PR) |
-| OpenAPI / `swagger.yaml` | Out of scope |
-| Queue API reference | Replace “Seedance consent” blurb with faces-off note + link to Seedance guide (all locales) |
-| Locales | Full i18n: EN + all locale copies of guide + queue blurb + `docs.json` nav |
-| Enhanced / Mini / 1.5 Pro | Out of this pass |
-
-## Faces-off messaging (required)
-
-Public docs must state:
-
-- Public API Seedance does **not** run face-asset / consent attestation (`consents.seedance` / `needs_consent` reseller path).
-- Face-bearing media may fail upstream (content-policy / provider rejection).
-- Likeness / face-asset generation: use the Venice app / Studio.
-- Do **not** link `seedance-face-consent` as a how-to for API callers.
-
-## Guide content changes
-
-File: `guides/media/seedance-2-0.mdx` (+ `ar|de|es|fr|it|ko|pt-BR|zh` copies)
-
-1. Title / description → Seedance 2.0 **and** 2.5.
-2. Variants table: keep 2.0 + Fast; add `seedance-2-5-{text,image,reference}-to-video`.
-3. Note R2V four-workflow model applies to 2.0 and 2.5 R2V.
-4. Prominent faces-off callout near the top (after intro / variants).
-5. **Split multimodal limits** by family (do not reuse a single 2.0 table for 2.5).
-
-### Limits (from outerface model defs)
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-| --- | --- | --- |
-| R2V reference images | 1–9 | 1–30 |
-| R2V reference videos | ≤3; combined ≤15s; per-clip 2–15s | ≤10; combined ≤30s; per-clip 2–30s |
-| R2V reference audio | ≤3; combined ≤15s | ≤10; combined ≤30s |
-| Output duration | 4–15s | 4–30s (default 10) |
-| Resolutions | 480p / 720p / 1080p (Fast: no 1080p) | 480p / 720p only |
-| Shared floors | short side ≥300px; aspect (0.4, 2.5); mime/size as today | same floors; max image bytes 30MB where defined |
-
-6. Keep most examples on 2.0 IDs; add ≥1 Seedance 2.5 example (longer duration and/or higher ref counts).
-
-## Nav
-
-`docs.json` — under Image, Video & Audio (every language), insert `…/guides/media/seedance-2-0` after `video-generation`.  
-Do **not** add `seedance-face-consent`.
-
-## API reference
-
-`api-reference/endpoint/video/queue.mdx` (+ locales): replace “### Seedance consent” with faces-off note linking to the Seedance guide. Keep “### Seedance 2.0” section; extend wording to 2.0 / 2.5 where it only names 2.0.
-
-## Out of scope
-
-- Rewriting or deleting `seedance-face-consent*`
-- `swagger.yaml` / OpenAPI sync
-- Video harness Seedance face-rule cleanup
-- Mini / Enhanced / 1.5 Pro public docs
-
-## Acceptance
-
-- No **nav-linked** public page instructs API callers to submit Seedance face consent.
-- Seedance guide documents faces-off + 2.0 vs 2.5 limits accurately.
-- Seedance guide is in public nav for all locales.
-- Face-consent pages remain off-nav with unchanged body.
-- Queue reference no longer points readers at the consent how-to as current API behavior.
diff --git a/es/api-reference/endpoint/video/queue.mdx b/es/api-reference/endpoint/video/queue.mdx
index 92410eba..32b0c08d 100644
--- a/es/api-reference/endpoint/video/queue.mdx
+++ b/es/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Llama a `/video/quote` para obtener una estimación de precio y, a continuación
 
 Los modelos privados también devuelven un `download_url` para el vídeo terminado. Es una URL de entrega de corta duración (algunos reintentos están bien si una descarga se interrumpe); consulta la [guía de generación de vídeo](/guides/media/video-generation#private-download-links) para conocer los detalles y el `DELETE` opcional para privacidad.
 
-### Seedance 2.0 y 2.5
+### Seedance 2.0
 
-Para los modelos `seedance-2-0-*-basic` y `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*` de Seedance 2.0), consulta la [Guía de Seedance 2.0 y 2.5](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), límites multimodales por familia, comportamiento faces-off de la API y detalles de precios.
+Para los modelos `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, además de las variantes `-fast-*`), consulta la [Guía de Seedance 2.0](/guides/media/seedance-2-0) para el modelo de cuatro flujos de trabajo (Reference / Edit / Extend / Stitch), patrones canónicos de prompt, límites de entrada multimodal y detalles de precios.
 
-### Seedance API sin rostros (faces-off)
+### Consentimiento de Seedance
 
-Los modelos Seedance públicos **no** admiten aprovisionamiento de face assets ni atestación de consentimiento facial. Las entradas con rostros pueden ser rechazadas upstream. Para generación con likeness / face assets, usa la app o Studio de Venice. Consulta la [guía de Seedance](/guides/media/seedance-2-0).
+Cuando una solicitud de image- o reference-to-video de Seedance contiene un rostro humano, la API devuelve un `409 needs_consent` con el texto de la política, y debes reenviar la misma solicitud con una atestación `consents.seedance`. Consulta la [Guía de consentimiento para contenido de rostro de Seedance](/guides/media/seedance-face-consent) para conocer el flujo completo, el objeto de consentimiento, deduplicación y revocación.
 
 ### Escalado de vídeo
 
diff --git a/es/guides/integrations/venice-video-harness.mdx b/es/guides/integrations/venice-video-harness.mdx
index 76d1b485..f4f03a51 100644
--- a/es/guides/integrations/venice-video-harness.mdx
+++ b/es/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Los defaults del harness son opinionados porque la consistencia es lo importante
 
 | Rol | Modelo predeterminado | Cuándo se usa |
 |---|---|---|
-| Tomas de personaje (1-2 personajes) | `seedance-2-0-reference-to-video-basic` | R2V por defecto con `reference_image_urls` planos, etiquetas `@Image`, hasta 15s, audio estéreo nativo |
+| Tomas de personaje (1-2 personajes) | `seedance-2-0-reference-to-video` | R2V por defecto con `reference_image_urls` planos, etiquetas `@Image`, hasta 15s, audio estéreo nativo |
 | Tomas de personaje (3+ personajes) | `kling-o3-standard-reference-to-video` | Fallback automático con `elements` estructurados para identidad multi-personaje |
-| Establecimiento / atmósfera / acción | `seedance-2-0-image-to-video-basic` | Sin personajes; calidad cinematográfica épica, hasta 15s |
+| Establecimiento / atmósfera / acción | `seedance-2-0-image-to-video` | Sin personajes; calidad cinematográfica épica, hasta 15s |
 
 Estos son sobrescribibles por proyecto vía `series.json → videoDefaults`. Para apuntar a una familia que no sea Seedance (p. ej., cuentas sin acceso a Seedance), establece `videoDefaults` en `kling-o3-standard-reference-to-video` y `veo3.1-fast-image-to-video`.
 
diff --git a/es/guides/media/seedance-2-0.mdx b/es/guides/media/seedance-2-0.mdx
index d202b63c..3d3b58ae 100644
--- a/es/guides/media/seedance-2-0.mdx
+++ b/es/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 y 2.5"
-description: "Genera, edita, amplía y une vídeos con Seedance 2.0 y 2.5 en Venice - flujos texto/imagen/referencia, comportamiento faces-off de la API y límites multimodales por familia."
-'og:title': "Guía Seedance 2.0 y 2.5 | Venice API Docs"
-'og:description': "Genera, edita, amplía y une vídeos con Seedance 2.0 y 2.5 vía la API de Venice. Patrones de flujo, faces-off, límites multimodales, precios y ejemplos."
+title: "Seedance 2.0"
+description: "Genera, edita, extiende y une vídeos con Seedance 2.0 en Venice — flujos de text, image y reference-to-video en un mismo endpoint, con ejemplos resueltos."
+'og:title': "Seedance 2.0 Guide | Venice API Docs"
+'og:description': "Genera, edita, extiende y une vídeos con Seedance 2.0 a través de la API de Venice. Patrones de flujo, límites multimodales, precios y ejemplos resueltos."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0 es un modelo multimodal de vídeo insignia expuesto en Venice como una familia de tres variantes para generación de vídeo a partir de texto, imagen y referencia. La variante **reference-to-video** es excepcionalmente potente: un único endpoint y un único ID de modelo gestionan **cuatro flujos distintos** (Reference, Edit, Extend, Stitch) — el flujo se infiere de la **forma de tu prompt**.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+Esta guía recorre las variantes, los cuatro flujos con sus prompts canónicos, los límites de entrada multimodal, los precios y ejemplos completos con `curl`.
 
-<Warning>
-**Los rostros no están soportados en la API pública de Seedance.** Venice no ejecuta aprovisionamiento de face assets ni atestación de consentimiento facial (`consents.seedance` / `needs_consent`) para los modelos Seedance públicos. Las entradas con rostros pueden rechazarse upstream como error de política de contenido / proveedor. Para generación con likeness / face assets, usa la app o Studio de Venice.
-</Warning>
-
-## Variants
+## Variantes
 
-| Model ID | Variant | Output resolutions | Notes |
+| ID del modelo | Variante | Resoluciones de salida | Notas |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Solo prompt de texto |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Anclaje de primer frame (y opcionalmente último) con imagen |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Hasta 9 imágenes de referencia + 3 vídeos de referencia + 3 donantes de audio. Impulsa Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Nivel más rápido y de menor fidelidad |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Nivel más rápido y de menor fidelidad |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Nivel más rápido y de menor fidelidad; mismo conjunto de flujos |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+Todas las variantes son asíncronas. Envía mediante `POST /api/v1/video/queue` y luego haz polling de `POST /api/v1/video/retrieve` hasta que el cuerpo de la respuesta sea `video/mp4`. Consulta [Generación de vídeo](/guides/media/video-generation) para el flujo general de cola.
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## El modelo "un modelo, cuatro flujos"
 
-## The "one model, four workflows" model
+La variante reference-to-video (`seedance-2-0-reference-to-video` y su hermana Fast) es el mismo modelo subyacente sirviendo cuatro tareas diferentes. **El modelo infiere la tarea a partir del prefijo del prompt y la forma de tus entradas.** No hay un campo `task` o `workflow` — la sintaxis del prompt es el enrutamiento.
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| Flujo | Qué hace | Prefijo del prompt | Entradas |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | Genera un vídeo nuevo usando los archivos de referencia subidos como donantes de sujeto / movimiento / estilo / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Texto + ≥1 referencia de imagen O vídeo (0-9 imágenes, 0-3 vídeos), más opcionalmente hasta 3 donantes de audio |
+| **Edit** | Modifica un único vídeo de entrada preservando el resto | `Strictly edit <Video 1>, changing its ...` | 1 vídeo de entrada + texto (imágenes como anclaje opcional) |
+| **Extend** | Extensión hacia adelante / atrás de un clip | `Extend <Video 1>, generate ...` | 1 vídeo de entrada + texto |
+| **Stitch** | Une 2-3 clips con transiciones auto-generadas | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 vídeos de entrada + texto |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+La **sintaxis del prompt es canónica y sensible a mayúsculas**: corchetes angulares, primera letra en mayúscula, un único espacio antes del número — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## Patrones de flujo
 
-### Reference workflow
+### Flujo Reference
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+Usa los archivos de referencia subidos como **donantes** — sujeto, escena, movimiento, estilo, timbre vocal — para generar un vídeo completamente nuevo.
 
-**Canonical prompt patterns**:
+**Patrones canónicos de prompt**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,23 +51,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**Ejemplos**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (los donantes de audio deben emparejarse con al menos una referencia de imagen o vídeo — solo audio se rechaza)
 
-### Edit workflow
+### Flujo Edit
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+Modifica un único vídeo de entrada. **Cualquier cosa no nombrada explícitamente en el prompt se preserva.** Úsalo cuando quieras un cambio localizado (cambio de sujeto, cambio de clima/color, añadir/eliminar elemento) en lugar de un vídeo totalmente nuevo.
 
-**Canonical prompt pattern**:
+**Patrón canónico de prompt**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**Sub-patrones para control más fino**:
 
 ```
 Add Elements:
@@ -89,136 +80,138 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**Ejemplos**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+El último ejemplo combina Edit con una referencia de imagen — totalmente legal, el modelo usa `<Image 1>` como donante visual para el reemplazo.
 
-### Extend workflow
+### Flujo Extend
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+Continúa un único clip hacia adelante o hacia atrás en el tiempo. **Por defecto, Seedance devuelve solo el contenido nuevo** — no el clip de entrada original concatenado con la extensión. Esto es por diseño, para continuidad de transición; si quieres preservar el clip de entrada junto con la extensión, dilo explícitamente:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserva el input al principio
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserva el input al final
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+Gestión de transiciones: el modelo extrae automáticamente los frames de transición para una mezcla sin costuras, y los segmentos originales del vídeo de entrada no se re-generan.
 
-**Examples**:
+**Ejemplos**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Flujo Stitch (Track Completion)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+Conecta 2-3 clips de entrada con transiciones generadas por IA. La duración total combinada de entrada debe ser ≤ 15 s.
 
-**Canonical prompt pattern**:
+**Patrón canónico de prompt**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**Ejemplos**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+El modelo recorta automáticamente los segmentos de conexión en los puntos de unión para la continuidad.
 
 ---
 
-## Universal prompt formula
+## Fórmula universal de prompt
 
-Across all four workflows, the recommended authoring formula is:
+En los cuatro flujos, la fórmula de autoría recomendada es:
 
 ```
-Subject + Motion + Environment (Optional)
-       + Camera Movement / Cut (Optional)
-       + Aesthetic Description (Optional)
-       + Audio (Optional)
+Sujeto + Movimiento + Entorno (opcional)
+       + Movimiento de cámara / Corte (opcional)
+       + Descripción estética (opcional)
+       + Audio (opcional)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Sujeto + Movimiento**: la base lógica — define "Quién" está realizando "Qué acción"
+- **Entorno + Estética**: fondo espacial, iluminación, estilo visual
+- **Cámara**: tipo o movimiento explícito de toma
+- **Audio**: efectos de sonido ambientales o dirección vocal para salida inmersiva
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+Poner esta capa encima de un prefijo de flujo (p. ej., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produce las salidas de mayor calidad.
 
 ---
 
-## Multimodal input limits
+## Límites de entrada multimodal
+
+Los valores siguientes son lo que acepta la API de Venice. Las solicitudes fuera de estos rangos se rechazan en la capa de schema con un 400 antes de llegar a la inferencia.
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### Imágenes
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| Restricción | Valor |
+|---|---|
+| Métodos de entrada | URL (`http://`, `https://`) o data URL en base64 (`data:image/...`) |
+| Formatos | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Aspect ratio (A / H) | exclusivo `(0.4, 2.5)` |
+| Lado mínimo | ≥ 300 px |
+| Cantidad de imágenes: primer frame I2V | 1 |
+| Cantidad de imágenes: primer + último frame I2V | 2 |
+| **Cantidad de imágenes: R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### Vídeos
 
-| Constraint | Value |
+| Restricción | Valor |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Métodos de entrada | URL (`http://`, `https://`) o data URL en base64 (`data:video/...`) |
+| Formatos | `.mp4`, `.mov` |
+| Códecs de vídeo | H.264 / AVC, H.265 / HEVC |
+| Códecs de audio (en el contenedor) | AAC, MP3 |
+| **Duración por clip** | `[2, 15]` s (inclusive) |
+| **Cantidad máx. de clips** | 3 (R2V / Stitch / Extend) |
+| **Duración combinada total** | ≤ 15 s entre todos los clips |
+| **Tamaño por clip** | ≤ 50 MB |
+
+### Audio
+
+| Restricción | Valor |
+|---|---|
+| Métodos de entrada | URL (`http://`, `https://`) o data URL en base64 (`data:audio/...`) |
+| Formatos | `.wav`, `.mp3` |
+| **Duración por clip** | `[2, 15]` s |
+| **Cantidad máx. de clips** | 3 |
+| **Duración combinada total** | ≤ 15 s entre todos los clips |
+| **Tamaño por clip** | ≤ 15 MB |
+
+El audio de referencia se admite solo en las variantes R2V. Cada entrada se reenvía al modelo como un elemento de contenido `role: "reference_audio"` al que el prompt se dirige como `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — el modelo usa cada clip para timbre vocal, efectos de sonido o música de fondo dependiendo de cómo el prompt lo encuadre. El campo singular heredado `audio_url` mapea a la misma forma de contenido y ahora es equivalente a pasar un `reference_audio_urls` de un elemento.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls` no puede ser la única entrada de referencia.** El modelo requiere al menos una referencia de imagen o vídeo junto con cualquier donante de audio. Empareja `reference_audio_urls` con `reference_image_urls`, `reference_video_urls`, `image_url` o `video_url` — los envíos solo de audio se rechazan.
 </Warning>
 
-### Request size
+### Tamaño de la solicitud
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+El endpoint de cola acepta cuerpos JSON de hasta **35 MB**. Las data URLs inline para vídeos grandes pueden empujar más allá de esto — para Stitch multi-clip en particular, prefiere URLs en lugar de base64 inline.
 
 ---
 
-## Pricing
+## Precios
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+Llama a `POST /api/v1/video/quote` para obtener una cotización para una forma de solicitud dada antes de enviarla a `/video/queue`. El endpoint de cotización es la única fuente autoritativa; los detalles de precio pueden cambiar y no deberían cachearse ni duplicarse en el cliente.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+Cuando los vídeos de referencia forman parte de la solicitud, pasa también `reference_video_total_duration` (la suma de todas las duraciones de clips de referencia en segundos) para que la cotización coincida con lo que cobrará `/video/queue`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,9 +221,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## Ejemplos completos
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+Todos los ejemplos asumen que `VENICE_API_KEY` está establecida en el entorno.
 
 ### Text-to-video
 
@@ -239,7 +232,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### Image-to-video (primer frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video` (y su variante Fast) **no aceptan `aspect_ratio`** — el aspect ratio de salida se deriva automáticamente de las dimensiones de la imagen de entrada. Pasar el campo devuelve un 400 con *"This model does not support aspect_ratio"*. Usa las variantes T2V o R2V si necesitas control explícito de aspect ratio.
 </Note>
 
-### Reference workflow — subject donor
+### Flujo Reference — donante de sujeto
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Flujo Reference — donante de sujeto + audio
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Flujo Edit
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### Flujo Edit con anclaje por imagen
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### Extender hacia adelante
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +351,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### Polling de finalización
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+Tras cada envío a la cola, guarda el `queue_id` devuelto y haz polling de `/video/retrieve` hasta que el cuerpo de la respuesta sea `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+La respuesta es JSON (`{ "status": "queued" | "running" | "failed", ... }`) hasta que el trabajo se complete, momento en el que el cuerpo de la respuesta cambia a bytes `video/mp4`. Consulta [Generación de vídeo](/guides/media/video-generation) para el patrón completo de polling.
 
 ---
 
-## Troubleshooting
+## Resolución de problemas
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Los envíos de reference-to-video deben incluir al menos uno de `reference_image_urls`, `reference_video_urls`, `image_references` o `video_references`. La generación pura solo de texto no es un flujo R2V válido — usa `seedance-2-0-text-to-video` en su lugar. `reference_audio_urls` por sí solo no es suficiente (consulta la sección de Audio anterior).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+El modelo limita los vídeos de referencia a 3. Si necesitas más clips, haz un Stitch primero (3 → 1) y luego usa la salida como referencia para un seguimiento.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / agregado `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+La duración por clip es `[2, 15]` segundos **inclusive**; la suma entre todos los vídeos de referencia también está limitada a 15 segundos. Recorta los clips en el cliente antes del envío.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### El prompt enruta al flujo incorrecto
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+El flujo se infiere de la sintaxis del prompt. Errores de enrutamiento comunes:
 
-### Face / likeness inputs rejected
+- Querer **Extend** pero escribir `Refer to ...` → el modelo trata tu vídeo como *donante*, no como lienzo a continuar
+- Querer **Stitch** pero escribir `Refer to ...` → el modelo elige uno como donante, ignora los demás
+- Querer **Edit** pero escribir `Generate a video based on <Video 1>` → ambiguo; el modelo puede caer en Reference por defecto
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+Usa los prefijos canónicos exactamente como están escritos: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### La cotización no coincide con el importe en cola
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+Si incluiste un vídeo de referencia pero no pasaste `reference_video_total_duration` a `/video/quote`, la cotización y el importe en cola pueden diferir. Pasa siempre `reference_video_total_duration` (suma de todas las duraciones de clips de referencia, en segundos) cuando haya vídeos de referencia.
 
 ---
 
-## References
+## Referencias
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Endpoint de cola de vídeo de Venice: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Endpoint de cotización de Venice: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Guía complementaria: [Reference to Video](/guides/media/reference-to-video) (cubre Kling O3 + Grok Imagine R2V)
+- Guía complementaria: [Generación de vídeo](/guides/media/video-generation) (visión general de cola / polling)
diff --git a/es/guides/media/video-generation.mdx b/es/guides/media/video-generation.mdx
index 31d9d880..b4c89542 100644
--- a/es/guides/media/video-generation.mdx
+++ b/es/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtén el coste exacto antes de generar. Envía solo entradas de precio (`model`
 La cotización es en USD.
 
 <Info>
-**Los parámetros de cotización son solo entradas de precio.** Los campos que envías a `/video/quote` —incluyendo `aspect_ratio`, `audio` y `reference_video_total_duration`— se usan únicamente para calcular el precio. **No** se reenvían a `/video/queue` y no afectan al vídeo generado. Enviar `aspect_ratio` a quote se acepta incluso para modelos (como `seedance-2-0-image-to-video-basic`) que rechazan `aspect_ratio` en el momento de la generación, ya que los modelos image-to-video derivan la relación de aspecto de salida a partir de la imagen de entrada.
+**Los parámetros de cotización son solo entradas de precio.** Los campos que envías a `/video/quote` —incluyendo `aspect_ratio`, `audio` y `reference_video_total_duration`— se usan únicamente para calcular el precio. **No** se reenvían a `/video/queue` y no afectan al vídeo generado. Enviar `aspect_ratio` a quote se acepta incluso para modelos (como `seedance-2-0-image-to-video`) que rechazan `aspect_ratio` en el momento de la generación, ya que los modelos image-to-video derivan la relación de aspecto de salida a partir de la imagen de entrada.
 </Info>
 
 ### Precios de reference-to-video
diff --git a/fr/api-reference/endpoint/video/queue.mdx b/fr/api-reference/endpoint/video/queue.mdx
index bae0b72b..e6902bb8 100644
--- a/fr/api-reference/endpoint/video/queue.mdx
+++ b/fr/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Appelez `/video/quote` pour obtenir une estimation du prix, puis interrogez `/vi
 
 Les modèles privés renvoient également un `download_url` pour la vidéo terminée. Il s'agit d'une URL de livraison à durée de vie courte (quelques relances sont acceptables si un téléchargement échoue) ; voir le [guide de génération vidéo](/guides/media/video-generation#private-download-links) pour les détails et la méthode `DELETE` optionnelle pour la confidentialité.
 
-### Seedance 2.0 et 2.5
+### Seedance 2.0
 
-Pour les modèles `seedance-2-0-*-basic` et `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*` de Seedance 2.0), voir le [Guide Seedance 2.0 et 2.5](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les limites multimodales par famille, le comportement faces-off de l’API et les tarifs.
+Pour les modèles `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, plus les variantes `-fast-*`), voir le [guide Seedance 2.0](/guides/media/seedance-2-0) pour le modèle à quatre workflows (Reference / Edit / Extend / Stitch), les modèles de prompt canoniques, les limites d'entrée multimodale et les détails de tarification.
 
-### API Seedance faces-off
+### Consentement Seedance
 
-Les modèles Seedance publics **ne** prennent **pas** en charge le provisioning d’assets de visage ni l’attestation de consentement. Les entrées comportant un visage peuvent être rejetées en amont. Pour la génération likeness / face assets, utilisez l’app ou Studio Venice. Voir le [guide Seedance](/guides/media/seedance-2-0).
+Lorsqu'une requête Seedance image-to-video ou reference-to-video contient un visage humain, l'API renvoie un `409 needs_consent` avec le texte de la politique, et vous resoumettez la même requête avec une attestation `consents.seedance`. Voir le [guide de consentement pour les médias avec visage Seedance](/guides/media/seedance-face-consent) pour le flux complet, l'objet de consentement, la déduplication et la révocation.
 
 ### Agrandissement vidéo
 
diff --git a/fr/guides/integrations/venice-video-harness.mdx b/fr/guides/integrations/venice-video-harness.mdx
index ff88276f..e0848ebc 100644
--- a/fr/guides/integrations/venice-video-harness.mdx
+++ b/fr/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Les défauts du harness sont assumés parce que la cohérence est le but. Le rou
 
 | Rôle | Modèle par défaut | Quand utilisé |
 |---|---|---|
-| Plans de personnage (1-2 personnages) | `seedance-2-0-reference-to-video-basic` | R2V par défaut avec `reference_image_urls` plats, tags `@Image`, jusqu'à 15 s, audio stéréo natif |
+| Plans de personnage (1-2 personnages) | `seedance-2-0-reference-to-video` | R2V par défaut avec `reference_image_urls` plats, tags `@Image`, jusqu'à 15 s, audio stéréo natif |
 | Plans de personnage (3+ personnages) | `kling-o3-standard-reference-to-video` | Fallback automatique avec `elements` structurés pour l'identité multi-personnages |
-| Plan d'établissement / d'ambiance / d'action | `seedance-2-0-image-to-video-basic` | Pas de personnages ; qualité cinématique épique, jusqu'à 15 s |
+| Plan d'établissement / d'ambiance / d'action | `seedance-2-0-image-to-video` | Pas de personnages ; qualité cinématique épique, jusqu'à 15 s |
 
 Ces valeurs sont remplaçables par projet via `series.json → videoDefaults`. Pour cibler une famille non-Seedance (par ex. des comptes qui n'ont pas accès à Seedance), définissez `videoDefaults` à `kling-o3-standard-reference-to-video` et `veo3.1-fast-image-to-video`.
 
diff --git a/fr/guides/media/seedance-2-0.mdx b/fr/guides/media/seedance-2-0.mdx
index e88d52d1..ace35ba7 100644
--- a/fr/guides/media/seedance-2-0.mdx
+++ b/fr/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 et 2.5"
-description: "Générez, modifiez, prolongez et assemblez des vidéos avec Seedance 2.0 et 2.5 sur Venice - workflows texte/image/référence, API faces-off et limites multimodales par famille."
-'og:title': "Guide Seedance 2.0 et 2.5 | Venice API Docs"
-'og:description': "Générez, modifiez, prolongez et assemblez des vidéos avec Seedance 2.0 et 2.5 via l'API Venice. Workflows, faces-off, limites, tarifs et exemples."
+title: "Seedance 2.0"
+description: "Générez, éditez, prolongez et assemblez des vidéos avec Seedance 2.0 sur Venice : workflows text, image et reference-to-video sur un seul endpoint."
+'og:title': "Guide Seedance 2.0 | Documentation de l'API Venice"
+'og:description': "Générez, éditez, prolongez et assemblez des vidéos avec Seedance 2.0 via l'API Venice. Modèles de workflow, limites multimodales, tarification et exemples concrets."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0 est un modèle vidéo multimodal phare exposé sur Venice sous forme d'une famille de trois variantes pour la génération vidéo pilotée par texte, image et référence. La variante **reference-to-video** est exceptionnellement puissante : un seul endpoint et un seul identifiant de modèle gèrent **quatre workflows distincts** (Reference, Edit, Extend, Stitch) — le workflow est inféré à partir de la **forme de votre prompt**.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+Ce guide passe en revue les variantes, les quatre workflows avec leurs prompts canoniques, les limites d'entrée multimodale, la tarification et des exemples `curl` complets.
 
-<Warning>
-**Les visages ne sont pas pris en charge sur l'API Seedance publique.** Venice n'exécute pas le provisioning d'assets de visage ni l'attestation de consentement (`consents.seedance` / `needs_consent`) pour les modèles Seedance publics. Les entrées comportant un visage peuvent être rejetées en amont (politique de contenu / fournisseur). Pour la génération likeness / face assets, utilisez l'app ou Studio Venice.
-</Warning>
-
-## Variants
+## Variantes
 
-| Model ID | Variant | Output resolutions | Notes |
+| Identifiant du modèle | Variante | Résolutions de sortie | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Prompt textuel uniquement |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Ancrage par première image (et optionnellement dernière image) |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Jusqu'à 9 images de référence + 3 vidéos de référence + 3 audios de référence. Alimente Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Palier plus rapide, fidélité moindre |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Palier plus rapide, fidélité moindre |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Palier plus rapide, fidélité moindre ; même ensemble de workflows |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+Toutes les variantes sont asynchrones. Soumettez via `POST /api/v1/video/queue`, puis interrogez `POST /api/v1/video/retrieve` jusqu'à ce que le corps de la réponse soit `video/mp4`. Voir [Génération vidéo](/guides/media/video-generation) pour le flux général de file d'attente.
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## Le modèle « un modèle, quatre workflows »
 
-## The "one model, four workflows" model
+La variante reference-to-video (`seedance-2-0-reference-to-video` et son équivalent Fast) est le même modèle sous-jacent servant quatre tâches différentes. **Le modèle déduit la tâche à partir du préfixe du prompt et de la forme de vos entrées.** Il n'y a pas de champ `task` ou `workflow` — la syntaxe du prompt est le routage.
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| Workflow | Ce qu'il fait | Préfixe de prompt | Entrées |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | Génère une nouvelle vidéo en utilisant les fichiers de référence téléversés comme donneurs pour le sujet / le mouvement / le style / l'audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Texte + ≥1 référence image OU vidéo (0-9 images, 0-3 vidéos), plus optionnellement jusqu'à 3 donneurs audio |
+| **Edit** | Modifie une seule vidéo d'entrée tout en préservant le reste | `Strictly edit <Video 1>, changing its ...` | 1 vidéo d'entrée + texte (ancrage par images optionnel) |
+| **Extend** | Prolongation avant / arrière d'un clip | `Extend <Video 1>, generate ...` | 1 vidéo d'entrée + texte |
+| **Stitch** | Assemble 2-3 clips avec des transitions générées automatiquement | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 vidéos d'entrée + texte |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+La **syntaxe du prompt est canonique et sensible à la casse** : crochets pointus, première lettre en majuscule, espace unique avant le numéro — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## Modèles de workflow
 
-### Reference workflow
+### Workflow Reference
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+Utilise les fichiers de référence téléversés comme **donneurs** — sujet, scène, mouvement, style, timbre vocal — pour générer une toute nouvelle vidéo.
 
-**Canonical prompt patterns**:
+**Modèles de prompt canoniques** :
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,23 +51,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**Exemples** :
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (les donneurs audio doivent être associés à au moins une référence image ou vidéo — l'audio seul est rejeté)
 
-### Edit workflow
+### Workflow Edit
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+Modifie une seule vidéo d'entrée. **Tout ce qui n'est pas explicitement nommé dans le prompt est préservé.** Utilisez-le lorsque vous voulez un changement localisé (remplacement de sujet, changement de météo/couleur, ajout/suppression d'élément) plutôt qu'une vidéo entièrement nouvelle.
 
-**Canonical prompt pattern**:
+**Modèle de prompt canonique** :
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**Sous-modèles pour un contrôle plus fin** :
 
 ```
 Add Elements:
@@ -89,57 +80,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**Exemples** :
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+Le dernier exemple combine Edit avec une référence d'image — parfaitement légal, le modèle utilise `<Image 1>` comme donneur visuel pour le remplacement.
 
-### Extend workflow
+### Workflow Extend
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+Prolonge un seul clip vers l'avant ou l'arrière dans le temps. **Par défaut, Seedance ne renvoie que le nouveau contenu** — pas l'entrée originale concaténée avec l'extension. C'est intentionnel, pour la continuité de transition ; si vous voulez que le clip d'entrée soit préservé à côté de l'extension, indiquez-le explicitement :
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← préserve l'entrée au début
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← préserve l'entrée à la fin
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+Gestion des transitions : le modèle extrait automatiquement les images de transition pour un fondu fluide, et les segments originaux de la vidéo d'entrée ne sont pas régénérés.
 
-**Examples**:
+**Exemples** :
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Workflow Stitch (Track Completion)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+Connecte 2-3 clips d'entrée avec des transitions générées par IA. La durée totale d'entrée combinée doit être ≤ 15 s.
 
-**Canonical prompt pattern**:
+**Modèle de prompt canonique** :
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**Exemples** :
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+Le modèle ajuste automatiquement les segments de liaison aux points de jonction pour assurer la continuité.
 
 ---
 
-## Universal prompt formula
+## Formule universelle de prompt
 
-Across all four workflows, the recommended authoring formula is:
+À travers les quatre workflows, la formule de rédaction recommandée est :
 
 ```
 Subject + Motion + Environment (Optional)
@@ -148,77 +139,79 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subject + Motion** : la base logique — définit « qui » effectue « quelle action »
+- **Environment + Aesthetics** : arrière-plan spatial, éclairage, style visuel
+- **Camera** : type ou mouvement de prise de vue explicite
+- **Audio** : effets sonores ambiants ou direction vocale pour une sortie immersive
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+Superposer cela à un préfixe de workflow (par exemple, `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produit les sorties de plus haute qualité.
 
 ---
 
-## Multimodal input limits
+## Limites d'entrée multimodale
+
+Les valeurs ci-dessous sont celles que l'API Venice accepte. Les requêtes hors de ces plages sont rejetées au niveau du schéma avec un 400 avant d'atteindre l'inférence.
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### Images
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| Contrainte | Valeur |
+|---|---|
+| Méthodes d'entrée | URL (`http://`, `https://`) ou URL data Base64 (`data:image/...`) |
+| Formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Ratio d'aspect (L / H) | exclusif `(0.4, 2.5)` |
+| Côté minimum | ≥ 300 px |
+| Nombre d'images : première image I2V | 1 |
+| Nombre d'images : première + dernière image I2V | 2 |
+| **Nombre d'images : R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### Vidéos
 
-| Constraint | Value |
+| Contrainte | Valeur |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Méthodes d'entrée | URL (`http://`, `https://`) ou URL data Base64 (`data:video/...`) |
+| Formats | `.mp4`, `.mov` |
+| Codecs vidéo | H.264 / AVC, H.265 / HEVC |
+| Codecs audio (dans le conteneur) | AAC, MP3 |
+| **Durée par clip** | `[2, 15]` s (inclusif) |
+| **Nombre maximum de clips** | 3 (R2V / Stitch / Extend) |
+| **Durée totale combinée** | ≤ 15 s à travers tous les clips |
+| **Taille par clip** | ≤ 50 Mo |
+
+### Audio
+
+| Contrainte | Valeur |
+|---|---|
+| Méthodes d'entrée | URL (`http://`, `https://`) ou URL data Base64 (`data:audio/...`) |
+| Formats | `.wav`, `.mp3` |
+| **Durée par clip** | `[2, 15]` s |
+| **Nombre maximum de clips** | 3 |
+| **Durée totale combinée** | ≤ 15 s à travers tous les clips |
+| **Taille par clip** | ≤ 15 Mo |
+
+L'audio de référence n'est pris en charge que sur les variantes R2V. Chaque entrée est transmise au modèle en tant qu'élément de contenu `role: "reference_audio"` que le prompt adresse comme `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — le modèle utilise chaque clip pour le timbre vocal, les effets sonores ou la musique de fond selon la façon dont le prompt le formule. L'ancien champ singulier `audio_url` mappe vers la même forme de contenu et équivaut désormais au passage d'un `reference_audio_urls` à un seul élément.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls` ne peut pas être la seule entrée de référence.** Le modèle requiert au moins une référence image ou vidéo en plus de tout donneur audio. Associez `reference_audio_urls` à `reference_image_urls`, `reference_video_urls`, `image_url` ou `video_url` — les soumissions audio seules sont rejetées.
 </Warning>
 
-### Request size
+### Taille de la requête
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+L'endpoint de file d'attente accepte des corps JSON jusqu'à **35 Mo**. Les URL data en ligne pour de grandes vidéos peuvent dépasser cette limite — pour le Stitch multi-clips en particulier, préférez les URL aux base64 en ligne.
 
 ---
 
-## Pricing
+## Tarification
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+Appelez `POST /api/v1/video/quote` pour obtenir un devis pour une forme de requête donnée avant de la soumettre à `/video/queue`. L'endpoint de devis est la seule source faisant autorité ; les détails de tarification peuvent changer et ne doivent pas être mis en cache ou dupliqués côté client.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+Lorsque des vidéos de référence font partie de la requête, transmettez également `reference_video_total_duration` (la somme de toutes les durées des clips de référence en secondes) afin que le devis corresponde à ce que `/video/queue` facturera :
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,9 +221,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## Exemples complets
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+Tous les exemples supposent que `VENICE_API_KEY` est défini dans l'environnement.
 
 ### Text-to-video
 
@@ -239,7 +232,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### Image-to-video (première image)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video` (et sa variante Fast) **n'accepte pas `aspect_ratio`** — le ratio d'aspect de sortie est dérivé automatiquement des dimensions de l'image d'entrée. Passer ce champ renvoie un 400 avec *« This model does not support aspect_ratio »*. Utilisez les variantes T2V ou R2V si vous avez besoin d'un contrôle explicite du ratio d'aspect.
 </Note>
 
-### Reference workflow — subject donor
+### Workflow Reference — donneur de sujet
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Workflow Reference — donneur sujet + audio
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Workflow Edit
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### Workflow Edit avec ancrage par image
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### Extend vers l'avant
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -407,7 +351,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### Sondage jusqu'à la fin
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+Après chaque soumission de file d'attente, enregistrez le `queue_id` retourné et interrogez `/video/retrieve` jusqu'à ce que le corps de la réponse soit `video/mp4` :
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+La réponse est en JSON (`{ "status": "queued" | "running" | "failed", ... }`) jusqu'à ce que la tâche se termine, moment auquel le corps de la réponse bascule vers des octets `video/mp4`. Voir [Génération vidéo](/guides/media/video-generation) pour le modèle complet de sondage.
 
 ---
 
-## Troubleshooting
+## Dépannage
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Les soumissions reference-to-video doivent inclure au moins l'un de `reference_image_urls`, `reference_video_urls`, `image_references` ou `video_references`. Une génération purement textuelle n'est pas un workflow R2V valide — utilisez plutôt `seedance-2-0-text-to-video`. `reference_audio_urls` seul ne suffit pas (voir la section Audio ci-dessus).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+Le modèle plafonne les vidéos de référence à 3. Si vous avez besoin de plus de clips, exécutez d'abord un Stitch (3 → 1), puis utilisez la sortie comme référence pour une suite.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / agrégat `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+La durée par clip est `[2, 15]` secondes **inclusif** ; la somme à travers toutes les vidéos de référence est également plafonnée à 15 secondes. Coupez les clips côté client avant la soumission.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### Le prompt achemine vers le mauvais workflow
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Le workflow est inféré à partir de la syntaxe du prompt. Erreurs d'acheminement courantes :
 
-### Face / likeness inputs rejected
+- Vouloir **Extend** mais écrire `Refer to ...` → le modèle traite votre vidéo comme un *donneur*, pas comme une toile à continuer
+- Vouloir **Stitch** mais écrire `Refer to ...` → le modèle en choisit une comme donneur, ignore les autres
+- Vouloir **Edit** mais écrire `Generate a video based on <Video 1>` → ambigu ; le modèle peut basculer par défaut sur Reference
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+Utilisez les préfixes canoniques exactement tels qu'écrits : `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### Le devis ne correspond pas au montant en file d'attente
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+Si vous avez inclus une vidéo de référence mais n'avez pas transmis `reference_video_total_duration` à `/video/quote`, le devis et le montant en file d'attente peuvent différer. Transmettez toujours `reference_video_total_duration` (somme de toutes les durées des clips de référence, en secondes) lorsque des vidéos de référence sont présentes.
 
 ---
 
-## References
+## Références
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Endpoint de file d'attente vidéo Venice : [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Endpoint de devis Venice : [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Guide complémentaire : [Reference to Video](/guides/media/reference-to-video) (couvre Kling O3 + Grok Imagine R2V)
+- Guide complémentaire : [Génération vidéo](/guides/media/video-generation) (vue d'ensemble queue / polling)
diff --git a/fr/guides/media/video-generation.mdx b/fr/guides/media/video-generation.mdx
index 5474cad8..6f0aa09f 100644
--- a/fr/guides/media/video-generation.mdx
+++ b/fr/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtenez le coût exact avant la génération. Envoyez uniquement les entrées de
 Le devis est en USD.
 
 <Info>
-**Les paramètres du devis servent uniquement au calcul du prix.** Les champs que vous envoyez à `/video/quote` — y compris `aspect_ratio`, `audio` et `reference_video_total_duration` — sont utilisés uniquement pour calculer le prix. Ils **ne sont pas** transmis à `/video/queue` et n'influencent pas la vidéo générée. Envoyer `aspect_ratio` au quote est accepté même pour les modèles (tels que `seedance-2-0-image-to-video-basic`) qui rejettent `aspect_ratio` au moment de la génération, puisque les modèles image-to-video dérivent le ratio d'aspect de sortie à partir de l'image d'entrée.
+**Les paramètres du devis servent uniquement au calcul du prix.** Les champs que vous envoyez à `/video/quote` — y compris `aspect_ratio`, `audio` et `reference_video_total_duration` — sont utilisés uniquement pour calculer le prix. Ils **ne sont pas** transmis à `/video/queue` et n'influencent pas la vidéo générée. Envoyer `aspect_ratio` au quote est accepté même pour les modèles (tels que `seedance-2-0-image-to-video`) qui rejettent `aspect_ratio` au moment de la génération, puisque les modèles image-to-video dérivent le ratio d'aspect de sortie à partir de l'image d'entrée.
 </Info>
 
 ### Tarification reference-to-video
diff --git a/it/api-reference/endpoint/video/queue.mdx b/it/api-reference/endpoint/video/queue.mdx
index b3fc9278..c65b84ad 100644
--- a/it/api-reference/endpoint/video/queue.mdx
+++ b/it/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Chiama `/video/quote` per ottenere una stima del prezzo, quindi esegui il pollin
 
 I modelli privati restituiscono anche un `download_url` per il video finito. Si tratta di un URL di consegna a breve termine (alcuni retry vanno bene se il download cade); consulta la [guida alla generazione video](/guides/media/video-generation#private-download-links) per i dettagli e l'uso opzionale di `DELETE` per la privacy.
 
-### Seedance 2.0 e 2.5
+### Seedance 2.0
 
-Per i modelli `seedance-2-0-*-basic` e `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*` di Seedance 2.0), vedi la [Guida Seedance 2.0 e 2.5](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i limiti multimodali per famiglia, il comportamento faces-off dell’API e i dettagli sui prezzi.
+Per i modelli `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, oltre alle varianti `-fast-*`), consulta la [guida a Seedance 2.0](/guides/media/seedance-2-0) per il modello a quattro workflow (Reference / Edit / Extend / Stitch), i pattern di prompt canonici, i limiti di input multimodali e i dettagli sui prezzi.
 
-### API Seedance faces-off
+### Consenso Seedance
 
-I modelli Seedance pubblici **non** supportano il provisioning di face asset né l’attestazione di consenso sul volto. Gli input con volti possono essere rifiutati upstream. Per la generazione likeness / face asset usa l’app o Studio Venice. Vedi la [guida Seedance](/guides/media/seedance-2-0).
+Quando una richiesta image- o reference-to-video di Seedance contiene un volto umano, l'API restituisce un `409 needs_consent` con il testo della policy e tu reinvii la stessa richiesta con un'attestazione `consents.seedance`. Consulta la [guida al consenso per i contenuti facciali di Seedance](/guides/media/seedance-face-consent) per il flusso completo, l'oggetto di consenso, la deduplica e la revoca.
 
 ### Upscaling video
 
diff --git a/it/guides/integrations/venice-video-harness.mdx b/it/guides/integrations/venice-video-harness.mdx
index 7a69efb6..c700692e 100644
--- a/it/guides/integrations/venice-video-harness.mdx
+++ b/it/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ I default dell'harness sono opinionated perché la consistenza è il punto chiav
 
 | Ruolo | Modello di default | Quando usato |
 |---|---|---|
-| Inquadrature con personaggi (1-2 personaggi) | `seedance-2-0-reference-to-video-basic` | R2V di default con `reference_image_urls` flat, tag `@Image`, fino a 15s, audio stereo nativo |
+| Inquadrature con personaggi (1-2 personaggi) | `seedance-2-0-reference-to-video` | R2V di default con `reference_image_urls` flat, tag `@Image`, fino a 15s, audio stereo nativo |
 | Inquadrature con personaggi (3+ personaggi) | `kling-o3-standard-reference-to-video` | Fallback automatico con `elements` strutturati per identità multi-personaggio |
-| Establishing / mood / action | `seedance-2-0-image-to-video-basic` | Nessun personaggio; qualità cinematica epica, fino a 15s |
+| Establishing / mood / action | `seedance-2-0-image-to-video` | Nessun personaggio; qualità cinematica epica, fino a 15s |
 
 Questi sono sovrascrivibili per progetto tramite `series.json → videoDefaults`. Per puntare a una famiglia non Seedance (es. account che non hanno accesso a Seedance), imposta `videoDefaults` su `kling-o3-standard-reference-to-video` e `veo3.1-fast-image-to-video`.
 
diff --git a/it/guides/media/seedance-2-0.mdx b/it/guides/media/seedance-2-0.mdx
index 947767b5..188a9c4d 100644
--- a/it/guides/media/seedance-2-0.mdx
+++ b/it/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 e 2.5"
-description: "Genera, modifica, estendi e unisci video con Seedance 2.0 e 2.5 su Venice - workflow testo/immagine/riferimento, API faces-off e limiti multimodali per famiglia."
-'og:title': "Guida Seedance 2.0 e 2.5 | Venice API Docs"
-'og:description': "Genera, modifica, estendi e unisci video con Seedance 2.0 e 2.5 tramite l'API Venice. Workflow, faces-off, limiti, prezzi ed esempi."
+title: "Seedance 2.0"
+description: "Genera, modifica, estendi e cuci video con Seedance 2.0 su Venice — workflow text-, image- e reference-to-video su un unico endpoint, con esempi pratici."
+'og:title': "Seedance 2.0 Guide | Venice API Docs"
+'og:description': "Genera, modifica, estendi e cuci video con Seedance 2.0 tramite l'API Venice. Pattern di workflow, limiti multimodali, prezzi ed esempi completi."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0 è un modello video multimodale flagship esposto su Venice come famiglia di tre varianti per la generazione video text-, image- e reference-driven. La variante **reference-to-video** è straordinariamente potente: un singolo endpoint e un singolo ID di modello gestiscono **quattro workflow distinti** (Reference, Edit, Extend, Stitch) — il workflow viene dedotto dalla **forma del tuo prompt**.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+Questa guida illustra le varianti, i quattro workflow con i loro prompt canonici, i limiti di input multimodali, i prezzi e esempi `curl` completi.
 
-<Warning>
-**I volti non sono supportati sull'API Seedance pubblica.** Venice non esegue provisioning di face asset né attestazione di consenso (`consents.seedance` / `needs_consent`) per i modelli Seedance pubblici. Gli input con volti possono essere rifiutati upstream come errore di content policy / provider. Per la generazione likeness / face asset usa l'app o Studio Venice.
-</Warning>
-
-## Variants
+## Varianti
 
-| Model ID | Variant | Output resolutions | Notes |
+| Model ID | Variante | Risoluzioni di output | Note |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Solo prompt testuale |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Grounding su first-frame (e opzionalmente last-frame) |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Fino a 9 immagini di riferimento + 3 video di riferimento + 3 donor audio di riferimento. Alimenta Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Tier più veloce, fidelity inferiore |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Tier più veloce, fidelity inferiore |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Tier più veloce, fidelity inferiore; stesso set di workflow |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+Tutte le varianti sono asincrone. Invia tramite `POST /api/v1/video/queue`, poi fai polling su `POST /api/v1/video/retrieve` finché il corpo della risposta non è `video/mp4`. Vedi [Video Generation](/guides/media/video-generation) per il flusso generale di queue.
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## Il modello "one model, four workflows"
 
-## The "one model, four workflows" model
+La variante reference-to-video (`seedance-2-0-reference-to-video` e il suo fratello Fast) è lo stesso modello sottostante che serve quattro task diversi. **Il modello deduce il task dal prefisso del prompt e dalla forma dei tuoi input.** Non c'è un campo `task` o `workflow` — la sintassi del prompt è il routing.
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| Workflow | Cosa fa | Prefisso del prompt | Input |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | Genera un nuovo video usando i file di riferimento caricati come donor per soggetto / movimento / stile / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Testo + ≥1 immagine OR video di riferimento (0-9 immagini, 0-3 video), più opzionalmente fino a 3 donor audio |
+| **Edit** | Modifica un singolo video di input preservando il resto | `Strictly edit <Video 1>, changing its ...` | 1 video di input + testo (immagini opzionali per grounding) |
+| **Extend** | Estensione in avanti / all'indietro di una clip | `Extend <Video 1>, generate ...` | 1 video di input + testo |
+| **Stitch** | Unisci 2-3 clip con transizioni auto-generate | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 video di input + testo |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+La **sintassi del prompt è canonica e case-sensitive**: parentesi angolari, prima lettera maiuscola, singolo spazio prima del numero — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## Pattern di workflow
 
-### Reference workflow
+### Workflow Reference
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+Usa i file di riferimento caricati come **donor** — soggetto, scena, movimento, stile, timbro vocale — per generare un video completamente nuovo.
 
-**Canonical prompt patterns**:
+**Pattern canonici del prompt**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,165 +51,167 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**Esempi**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (i donor audio devono essere abbinati ad almeno un'immagine o video di riferimento — solo audio viene rifiutato)
 
-### Edit workflow
+### Workflow Edit
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+Modifica un singolo video di input. **Tutto ciò che non è esplicitamente nominato nel prompt viene preservato.** Usalo quando vuoi un cambiamento localizzato (sostituzione del soggetto, cambio di meteo/colore, aggiunta/rimozione di elementi) piuttosto che un video completamente nuovo.
 
-**Canonical prompt pattern**:
+**Pattern canonico del prompt**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**Sotto-pattern per un controllo più fine**:
 
 ```
-Add Elements:
+Aggiungere elementi:
   At [timestamp / timing] and [spatial location] of <Video 1>, add [description of intended element].
 
-Remove Elements:
+Rimuovere elementi:
   Remove [element to be deleted] from <Video 1>, keeping the rest of the video content unchanged.
 
-Modify Elements:
+Modificare elementi:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**Esempi**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+L'ultimo esempio combina Edit con un riferimento di immagine — è perfettamente legale, il modello usa `<Image 1>` come donor visivo per la sostituzione.
 
-### Extend workflow
+### Workflow Extend
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+Continua una singola clip in avanti o all'indietro nel tempo. **Di default Seedance restituisce solo il nuovo contenuto** — non l'input originale concatenato con l'estensione. Questo è di proposito, per la continuità della transizione; se vuoi che la clip di input sia preservata accanto all'estensione, dillo esplicitamente:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserva l'input all'inizio
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserva l'input alla fine
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+Gestione della transizione: il modello estrae automaticamente i frame di transizione per un blending senza soluzione di continuità, e i segmenti originali del video di input non vengono ri-generati.
 
-**Examples**:
+**Esempi**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Workflow Stitch (Track Completion)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+Collega 2-3 clip di input con transizioni generate dall'AI. La durata combinata totale di input deve essere ≤ 15 s.
 
-**Canonical prompt pattern**:
+**Pattern canonico del prompt**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**Esempi**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+Il modello taglia automaticamente i segmenti di connessione nei punti di giuntura per garantire continuità.
 
 ---
 
-## Universal prompt formula
+## Formula universale del prompt
 
-Across all four workflows, the recommended authoring formula is:
+In tutti e quattro i workflow, la formula di scrittura consigliata è:
 
 ```
-Subject + Motion + Environment (Optional)
-       + Camera Movement / Cut (Optional)
-       + Aesthetic Description (Optional)
-       + Audio (Optional)
+Subject + Motion + Environment (Opzionale)
+       + Camera Movement / Cut (Opzionale)
+       + Aesthetic Description (Opzionale)
+       + Audio (Opzionale)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subject + Motion**: la base logica — definisce "Chi" sta eseguendo "Quale azione"
+- **Environment + Aesthetics**: sfondo spaziale, illuminazione, stile visivo
+- **Camera**: tipo di inquadratura o movimento esplicito
+- **Audio**: effetti sonori ambientali o direzione vocale per un output immersivo
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+Sovrapporre questo a un prefisso di workflow (es. `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produce gli output di qualità più alta.
 
 ---
 
-## Multimodal input limits
+## Limiti di input multimodale
+
+I valori sottostanti sono quelli che l'API Venice accetta. Le richieste al di fuori di questi range vengono rifiutate a livello di schema con un 400 prima di raggiungere l'inferenza.
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### Immagini
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| Vincolo | Valore |
+|---|---|
+| Metodi di input | URL (`http://`, `https://`) o data URL Base64 (`data:image/...`) |
+| Formati | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Aspect ratio (W / H) | esclusivo `(0.4, 2.5)` |
+| Lato minimo | ≥ 300 px |
+| Conteggio immagini: I2V first-frame | 1 |
+| Conteggio immagini: I2V first + last frame | 2 |
+| **Conteggio immagini: R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### Video
 
-| Constraint | Value |
+| Vincolo | Valore |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Metodi di input | URL (`http://`, `https://`) o data URL Base64 (`data:video/...`) |
+| Formati | `.mp4`, `.mov` |
+| Codec video | H.264 / AVC, H.265 / HEVC |
+| Codec audio (nel container) | AAC, MP3 |
+| **Durata per clip** | `[2, 15]` s (inclusivo) |
+| **Numero massimo di clip** | 3 (R2V / Stitch / Extend) |
+| **Durata totale combinata** | ≤ 15 s tra tutte le clip |
+| **Dimensione per clip** | ≤ 50 MB |
+
+### Audio
+
+| Vincolo | Valore |
+|---|---|
+| Metodi di input | URL (`http://`, `https://`) o data URL Base64 (`data:audio/...`) |
+| Formati | `.wav`, `.mp3` |
+| **Durata per clip** | `[2, 15]` s |
+| **Numero massimo di clip** | 3 |
+| **Durata totale combinata** | ≤ 15 s tra tutte le clip |
+| **Dimensione per clip** | ≤ 15 MB |
+
+L'audio di riferimento è supportato solo sulle varianti R2V. Ogni entry viene inoltrata al modello come content item `role: "reference_audio"` che il prompt richiama come `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — il modello usa ciascuna clip per timbro vocale, effetti sonori o musica di sottofondo a seconda di come la inquadra il prompt. Il campo legacy singolare `audio_url` viene mappato sulla stessa forma di contenuto ed è ora equivalente a passare un `reference_audio_urls` di un solo elemento.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls` non può essere l'unico input di riferimento.** Il modello richiede almeno un'immagine o un video di riferimento accanto a qualsiasi donor audio. Abbina `reference_audio_urls` con `reference_image_urls`, `reference_video_urls`, `image_url` o `video_url` — le submission solo audio vengono rifiutate.
 </Warning>
 
-### Request size
+### Dimensione della richiesta
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+L'endpoint di queue accetta corpi JSON fino a **35 MB**. I data URL inline per video di grandi dimensioni possono superare questa soglia — per Stitch multi-clip in particolare, preferisci gli URL al base64 inline.
 
 ---
 
-## Pricing
+## Prezzi
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+Chiama `POST /api/v1/video/quote` per ottenere un preventivo per una determinata forma di richiesta prima di inviarla a `/video/queue`. L'endpoint quote è l'unica fonte autorevole; i dettagli sui prezzi possono cambiare e non dovrebbero essere messi in cache o duplicati lato client.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+Quando uno o più video di riferimento fanno parte della richiesta, passa anche `reference_video_total_duration` (la somma di tutte le durate delle clip di riferimento in secondi) in modo che il preventivo corrisponda a quello che `/video/queue` addebiterà:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,9 +221,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## Esempi completi
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+Tutti gli esempi assumono che `VENICE_API_KEY` sia impostata nell'environment.
 
 ### Text-to-video
 
@@ -239,7 +232,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,36 +240,6 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
 ### Image-to-video (first frame)
 
 ```bash
@@ -284,7 +247,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video` (e la sua variante Fast) **non accettano `aspect_ratio`** — l'aspect ratio di output viene auto-derivato dalle dimensioni dell'immagine di input. Passare il campo restituisce un 400 con *"This model does not support aspect_ratio"*. Usa le varianti T2V o R2V se hai bisogno di un controllo esplicito sull'aspect ratio.
 </Note>
 
-### Reference workflow — subject donor
+### Workflow Reference — subject donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Workflow Reference — subject + audio donor
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Workflow Edit
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### Workflow Edit con image grounding
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### Extend in avanti
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -400,14 +344,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clips)
+### Stitch (3 clip)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,22 +365,22 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### Polling per il completamento
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+Dopo ogni submission in queue, salva il `queue_id` restituito e fai polling su `/video/retrieve` finché il corpo della risposta non è `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+La risposta è JSON (`{ "status": "queued" | "running" | "failed", ... }`) finché il job non è completato, momento in cui il corpo della risposta passa a byte `video/mp4`. Vedi [Video Generation](/guides/media/video-generation) per il pattern di polling completo.
 
 ---
 
@@ -444,42 +388,35 @@ The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) unti
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Le submission reference-to-video devono includere almeno uno tra `reference_image_urls`, `reference_video_urls`, `image_references` o `video_references`. La generazione puramente solo-testo non è un workflow R2V valido — usa invece `seedance-2-0-text-to-video`. `reference_audio_urls` da solo non è sufficiente (vedi la sezione Audio sopra).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+Il modello limita i video di riferimento a 3. Se hai bisogno di più clip, esegui prima uno Stitch (3 → 1), poi usa l'output come riferimento per un follow-up.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / aggregato `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+La durata per clip è `[2, 15]` secondi **inclusivo**; la somma su tutti i video di riferimento è anche limitata a 15 secondi. Taglia le clip lato client prima della submission.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### Il prompt viene instradato al workflow sbagliato
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+Il workflow è dedotto dalla sintassi del prompt. Misroute comuni:
 
-### Face / likeness inputs rejected
+- Volere **Extend** ma scrivere `Refer to ...` → il modello tratta il tuo video come un *donor*, non come una canvas da continuare
+- Volere **Stitch** ma scrivere `Refer to ...` → il modello sceglie uno come donor e ignora gli altri
+- Volere **Edit** ma scrivere `Generate a video based on <Video 1>` → ambiguo; il modello potrebbe ripiegare su Reference
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+Usa i prefissi canonici esattamente come scritti: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### Il quote non corrisponde all'importo della queue
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+Se hai incluso un video di riferimento ma non hai passato `reference_video_total_duration` a `/video/quote`, il quote e l'importo della queue possono differire. Passa sempre `reference_video_total_duration` (somma di tutte le durate delle clip di riferimento, in secondi) quando sono presenti video di riferimento.
 
 ---
 
-## References
+## Riferimenti
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Endpoint Venice video queue: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Endpoint Venice quote: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Guida companion: [Reference to Video](/guides/media/reference-to-video) (copre Kling O3 + Grok Imagine R2V)
+- Guida companion: [Video Generation](/guides/media/video-generation) (panoramica queue / polling)
diff --git a/it/guides/media/video-generation.mdx b/it/guides/media/video-generation.mdx
index b820d7b1..b5d6da5b 100644
--- a/it/guides/media/video-generation.mdx
+++ b/it/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Ottieni il costo esatto prima di generare. Invia solo gli input di pricing (`mod
 Il preventivo è in USD.
 
 <Info>
-**I parametri della quote sono solo input di pricing.** I campi che invii a `/video/quote` — inclusi `aspect_ratio`, `audio` e `reference_video_total_duration` — vengono usati esclusivamente per calcolare il prezzo. **Non** vengono inoltrati a `/video/queue` e non influenzano il video generato. Inviare `aspect_ratio` alla quote è accettato anche per i modelli (come `seedance-2-0-image-to-video-basic`) che rifiutano `aspect_ratio` al momento della generazione, poiché i modelli image-to-video ricavano l'aspect ratio di output dall'immagine di input.
+**I parametri della quote sono solo input di pricing.** I campi che invii a `/video/quote` — inclusi `aspect_ratio`, `audio` e `reference_video_total_duration` — vengono usati esclusivamente per calcolare il prezzo. **Non** vengono inoltrati a `/video/queue` e non influenzano il video generato. Inviare `aspect_ratio` alla quote è accettato anche per i modelli (come `seedance-2-0-image-to-video`) che rifiutano `aspect_ratio` al momento della generazione, poiché i modelli image-to-video ricavano l'aspect ratio di output dall'immagine di input.
 </Info>
 
 ### Prezzi reference-to-video
diff --git a/ko/api-reference/endpoint/video/queue.mdx b/ko/api-reference/endpoint/video/queue.mdx
index 03007670..9ba663af 100644
--- a/ko/api-reference/endpoint/video/queue.mdx
+++ b/ko/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ openapi: 'POST /video/queue'
 
 비공개 모델은 완성된 비디오를 위한 `download_url`도 반환합니다. 이는 수명이 짧은 전달 URL입니다(다운로드가 끊겨도 몇 번의 재시도는 괜찮습니다). 자세한 내용과 개인정보 보호를 위한 선택적 `DELETE`는 [비디오 생성 가이드](/guides/media/video-generation#private-download-links)를 참조하세요.
 
-### Seedance 2.0 & 2.5
+### Seedance 2.0
 
-`seedance-2-0-*-basic` 및 `seedance-2-5-*-basic` 모델(text-to-video, image-to-video, reference-to-video, 그리고 Seedance 2.0 `-fast-*` 변형)은 [Seedance 2.0 & 2.5 가이드](/guides/media/seedance-2-0)에서 네 가지 워크플로(Reference / Edit / Extend / Stitch), 패밀리별 멀티모달 한도, faces-off API 동작, 가격 정보를 확인하세요.
+`seedance-2-0-*` 모델(text-to-video, image-to-video, reference-to-video 및 `-fast-*` 변형)의 경우, 네 가지 워크플로 모델(Reference / Edit / Extend / Stitch), 표준 프롬프트 패턴, 멀티모달 입력 제한 및 가격 세부 정보는 [Seedance 2.0 가이드](/guides/media/seedance-2-0)를 참조하세요.
 
-### Seedance faces-off API
+### Seedance 동의
 
-공개 Seedance 모델은 페이스 에셋 프로비저닝이나 얼굴 동의 증명을 **지원하지 않습니다**. 얼굴이 포함된 입력은 업스트림에서 거부될 수 있습니다. likeness / 페이스 에셋 생성은 Venice 앱 또는 Studio를 사용하세요. 자세한 내용은 [Seedance 가이드](/guides/media/seedance-2-0)를 참고하세요.
+Seedance image- 또는 reference-to-video 요청에 사람의 얼굴이 포함된 경우, API는 정책 텍스트와 함께 `409 needs_consent`를 반환하며, `consents.seedance` 증명과 함께 동일한 요청을 다시 제출합니다. 전체 흐름, 동의 객체, 중복 제거 및 철회에 대해서는 [Seedance 얼굴 미디어 동의 가이드](/guides/media/seedance-face-consent)를 참조하세요.
 
 ### 비디오 업스케일링
 
diff --git a/ko/guides/integrations/venice-video-harness.mdx b/ko/guides/integrations/venice-video-harness.mdx
index 7e50d081..3664067c 100644
--- a/ko/guides/integrations/venice-video-harness.mdx
+++ b/ko/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ curl -L -o ~/.cache/whisper.cpp/ggml-base.en.bin \
 
 | Role | Default model | When used |
 |---|---|---|
-| 캐릭터 샷(1-2 캐릭터) | `seedance-2-0-reference-to-video-basic` | 평면 `reference_image_urls`, `@Image` 태그, 최대 15초, 네이티브 스테레오 오디오의 기본 R2V |
+| 캐릭터 샷(1-2 캐릭터) | `seedance-2-0-reference-to-video` | 평면 `reference_image_urls`, `@Image` 태그, 최대 15초, 네이티브 스테레오 오디오의 기본 R2V |
 | 캐릭터 샷(3+ 캐릭터) | `kling-o3-standard-reference-to-video` | 멀티 캐릭터 정체성을 위해 구조화된 `elements`로 자동 fallback |
-| 설정 / 분위기 / 액션 | `seedance-2-0-image-to-video-basic` | 캐릭터 없음. 서사적 시네마틱 품질, 최대 15초 |
+| 설정 / 분위기 / 액션 | `seedance-2-0-image-to-video` | 캐릭터 없음. 서사적 시네마틱 품질, 최대 15초 |
 
 이 기본값은 `series.json → videoDefaults`로 프로젝트별로 덮어쓸 수 있습니다. Seedance가 아닌 패밀리(예: Seedance 접근 권한이 없는 계정)를 타겟팅하려면 `videoDefaults`를 `kling-o3-standard-reference-to-video`와 `veo3.1-fast-image-to-video`로 설정하세요.
 
diff --git a/ko/guides/media/seedance-2-0.mdx b/ko/guides/media/seedance-2-0.mdx
index c3c88435..83f1c878 100644
--- a/ko/guides/media/seedance-2-0.mdx
+++ b/ko/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 & 2.5"
-description: "Venice에서 Seedance 2.0 및 2.5로 영상을 생성·편집·확장·스티치 - 텍스트/이미지/레퍼런스 워크플로, faces-off API, 패밀리별 멀티모달 한도."
-'og:title': "Seedance 2.0 & 2.5 가이드 | Venice API Docs"
-'og:description': "Venice API로 Seedance 2.0 및 2.5 영상을 생성·편집·확장·스티치. 워크플로, faces-off, 멀티모달 한도, 가격, 예제."
+title: "Seedance 2.0"
+description: "Venice의 Seedance 2.0으로 비디오를 생성·편집·확장·스티치하세요 — 하나의 엔드포인트에서 텍스트, 이미지, 레퍼런스-투-비디오 워크플로를 실제 예제와 함께 제공합니다."
+'og:title': "Seedance 2.0 Guide | Venice API Docs"
+'og:description': "Venice API를 통해 Seedance 2.0으로 비디오를 생성, 편집, 확장, 스티치하세요. 워크플로 패턴, 멀티모달 한도, 가격, 실제 예시."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0은 텍스트, 이미지, 레퍼런스 기반 비디오 생성을 위해 세 가지 변형 패밀리로 Venice에 노출된 플래그십 멀티모달 비디오 모델입니다. **reference-to-video** 변형은 특히 강력합니다: 단일 endpoint와 단일 모델 ID가 **네 가지 구분된 워크플로**(Reference, Edit, Extend, Stitch)를 처리합니다 — 워크플로는 **prompt 형태**에서 추론됩니다.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+이 가이드는 변형들, 표준 prompt와 함께 네 가지 워크플로, 멀티모달 input 한도, 가격, 완전한 `curl` 예시를 다룹니다.
 
-<Warning>
-**공개 Seedance API에서는 얼굴을 지원하지 않습니다.** Venice는 공개 Seedance 모델에 대해 페이스 에셋 프로비저닝이나 얼굴 동의 증명(`consents.seedance` / `needs_consent`)을 실행하지 않습니다. 얼굴이 포함된 입력은 업스트림에서 콘텐츠 정책 / 제공자 오류로 거부될 수 있습니다. likeness / 페이스 에셋 생성은 Venice 앱 또는 Studio를 사용하세요.
-</Warning>
-
-## Variants
+## 변형(Variants)
 
 | Model ID | Variant | Output resolutions | Notes |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
-
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | 텍스트 prompt만 |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | First-frame(과 선택적으로 last-frame) 이미지 그라운딩 |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | 최대 9장의 레퍼런스 이미지 + 3개의 레퍼런스 비디오 + 3개의 레퍼런스 오디오 도너. Reference / Edit / Extend / Stitch 구동 |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | 더 빠르고, 더 낮은 충실도 등급 |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | 더 빠르고, 더 낮은 충실도 등급 |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | 더 빠르고, 더 낮은 충실도 등급. 동일한 워크플로 세트 |
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+모든 변형은 비동기입니다. `POST /api/v1/video/queue`로 제출한 다음, 응답 본문이 `video/mp4`가 될 때까지 `POST /api/v1/video/retrieve`로 폴링하세요. 일반적인 큐 흐름은 [Video Generation](/guides/media/video-generation)을 참고하세요.
 
-## The "one model, four workflows" model
+## "하나의 모델, 네 가지 워크플로" 모델
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
+reference-to-video 변형(`seedance-2-0-reference-to-video`와 그의 Fast 형제)은 네 가지 다른 작업을 처리하는 동일한 기반 모델입니다. **모델은 prompt prefix와 input 형태에서 작업을 추론합니다.** `task`나 `workflow` 필드가 없습니다 — prompt 문법이 라우팅입니다.
 
 | Workflow | What it does | Prompt prefix | Inputs |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | 업로드한 레퍼런스 파일을 피사체 / 모션 / 스타일 / 오디오의 도너로 사용해 새 비디오 생성 | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | 텍스트 + 최소 1개의 이미지 또는 비디오 레퍼런스(이미지 0-9개, 비디오 0-3개), 선택적으로 최대 3개 오디오 도너 |
+| **Edit** | 나머지를 보존하면서 하나의 input 비디오를 수정 | `Strictly edit <Video 1>, changing its ...` | 1개의 input 비디오 + 텍스트(이미지 선택적 그라운딩) |
+| **Extend** | 한 클립의 앞으로 / 뒤로 확장 | `Extend <Video 1>, generate ...` | 1개의 input 비디오 + 텍스트 |
+| **Stitch** | 자동 생성 전환으로 2-3개 클립 연결 | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3개의 input 비디오 + 텍스트 |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+**Prompt 문법은 표준이며 대소문자 구분합니다**: 꺾쇠 괄호, 대문자 첫 글자, 숫자 앞 단일 공백 — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## 워크플로 패턴
 
-### Reference workflow
+### Reference 워크플로
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+업로드된 레퍼런스 파일을 **도너** — 피사체, 장면, 모션, 스타일, 목소리 음색 — 로 사용해 완전히 새로운 비디오를 생성합니다.
 
-**Canonical prompt patterns**:
+**표준 prompt 패턴**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,165 +51,167 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**예시**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (오디오 도너는 최소 하나의 이미지 또는 비디오 레퍼런스와 짝지어져야 합니다 — 오디오 단독은 거부됩니다)
 
-### Edit workflow
+### Edit 워크플로
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+하나의 input 비디오를 수정합니다. **prompt에 명시되지 않은 것은 모두 보존됩니다.** 완전히 새로운 비디오가 아닌 국소적 변경(피사체 교체, 날씨/색상 변경, 요소 추가/제거)을 원할 때 사용하세요.
 
-**Canonical prompt pattern**:
+**표준 prompt 패턴**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**더 세밀한 제어를 위한 서브 패턴**:
 
 ```
-Add Elements:
+요소 추가:
   At [timestamp / timing] and [spatial location] of <Video 1>, add [description of intended element].
 
-Remove Elements:
+요소 제거:
   Remove [element to be deleted] from <Video 1>, keeping the rest of the video content unchanged.
 
-Modify Elements:
+요소 수정:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**예시**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+마지막 예시는 Edit를 이미지 레퍼런스와 결합합니다 — 완전히 적법하며, 모델은 `<Image 1>`을 교체 대상의 시각적 도너로 사용합니다.
 
-### Extend workflow
+### Extend 워크플로
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+한 클립을 시간상 앞으로 또는 뒤로 이어갑니다. **기본적으로 Seedance는 새 콘텐츠만 반환합니다** — 원본 input이 확장과 함께 이어진 형태가 아닙니다. 이는 전환 연속성을 위한 의도된 동작입니다. input 클립을 확장과 함께 보존하고 싶다면 명시적으로 그렇게 말하세요:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← 시작에 input 보존
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← 끝에 input 보존
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+전환 처리: 모델이 매끄러운 블렌딩을 위해 전환 프레임을 자동으로 추출하며, input 비디오의 원본 세그먼트는 재생성되지 않습니다.
 
-**Examples**:
+**예시**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Stitch 워크플로(Track Completion)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+2-3개의 input 클립을 AI 생성 전환으로 연결합니다. 결합된 input 총 길이는 15초 이하여야 합니다.
 
-**Canonical prompt pattern**:
+**표준 prompt 패턴**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**예시**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+모델은 연속성을 위해 접합점에서 연결 세그먼트를 자동으로 트림합니다.
 
 ---
 
-## Universal prompt formula
+## 범용 prompt 공식
 
-Across all four workflows, the recommended authoring formula is:
+네 가지 워크플로 모두에서 권장되는 작성 공식:
 
 ```
-Subject + Motion + Environment (Optional)
-       + Camera Movement / Cut (Optional)
-       + Aesthetic Description (Optional)
-       + Audio (Optional)
+Subject + Motion + Environment (선택)
+       + Camera Movement / Cut (선택)
+       + Aesthetic Description (선택)
+       + Audio (선택)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subject + Motion**: 논리적 기반 — "누가" "어떤 행동"을 하는지 정의
+- **Environment + Aesthetics**: 공간 배경, 조명, 시각 스타일
+- **Camera**: 명시적 샷 유형 또는 움직임
+- **Audio**: 몰입감 있는 출력을 위한 환경음 또는 보이스 디렉션
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+이를 워크플로 prefix(예: `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) 위에 레이어링하면 최고 품질의 출력을 만들어냅니다.
 
 ---
 
-## Multimodal input limits
+## 멀티모달 input 한도
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+아래 값은 Venice API가 허용하는 값입니다. 이 범위를 벗어나는 요청은 추론에 도달하기 전 스키마 레이어에서 400으로 거부됩니다.
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+### 이미지
+
+| Constraint | Value |
+|---|---|
+| Input 방식 | URL(`http://`, `https://`) 또는 Base64 data URL(`data:image/...`) |
+| 포맷 | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| 종횡비(W / H) | 배타적 `(0.4, 2.5)` |
+| 최소 변 | ≥ 300 px |
+| 이미지 수: I2V first-frame | 1 |
+| 이미지 수: I2V first + last frame | 2 |
+| **이미지 수: R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### 비디오
 
 | Constraint | Value |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Input 방식 | URL(`http://`, `https://`) 또는 Base64 data URL(`data:video/...`) |
+| 포맷 | `.mp4`, `.mov` |
+| 비디오 코덱 | H.264 / AVC, H.265 / HEVC |
+| 오디오 코덱(컨테이너 내) | AAC, MP3 |
+| **클립당 길이** | `[2, 15]` 초(포함) |
+| **최대 클립 수** | 3 (R2V / Stitch / Extend) |
+| **결합 총 길이** | 모든 클립에 걸쳐 ≤ 15초 |
+| **클립당 크기** | ≤ 50 MB |
+
+### 오디오
+
+| Constraint | Value |
+|---|---|
+| Input 방식 | URL(`http://`, `https://`) 또는 Base64 data URL(`data:audio/...`) |
+| 포맷 | `.wav`, `.mp3` |
+| **클립당 길이** | `[2, 15]` 초 |
+| **최대 클립 수** | 3 |
+| **결합 총 길이** | 모든 클립에 걸쳐 ≤ 15초 |
+| **클립당 크기** | ≤ 15 MB |
+
+레퍼런스 오디오는 R2V 변형에서만 지원됩니다. 각 항목은 모델에 `role: "reference_audio"` 콘텐츠 항목으로 전달되며, prompt는 `<Audio 1>`, `<Audio 2>`, `<Audio 3>`으로 주소를 지정합니다 — 모델은 prompt가 그것을 어떻게 프레이밍하느냐에 따라 각 클립을 보이스 음색, 효과음, 배경 음악에 사용합니다. 레거시 단수 `audio_url` 필드는 같은 콘텐츠 형태로 매핑되며, 이제 한 개 요소 `reference_audio_urls`를 전달하는 것과 동등합니다.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls`만 단독으로 레퍼런스 input일 수는 없습니다.** 모델은 모든 오디오 도너와 함께 최소 하나의 이미지 또는 비디오 레퍼런스가 있어야 합니다. `reference_audio_urls`를 `reference_image_urls`, `reference_video_urls`, `image_url`, `video_url`과 짝지으세요 — 오디오 전용 제출은 거부됩니다.
 </Warning>
 
-### Request size
+### 요청 크기
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+queue endpoint는 최대 **35 MB**의 JSON 본문을 받습니다. 대용량 비디오의 인라인 data URL은 이를 초과할 수 있습니다 — 특히 멀티 클립 Stitch에서는 인라인 base64보다 URL을 선호하세요.
 
 ---
 
-## Pricing
+## 가격
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+`/video/queue`에 제출하기 전에 주어진 요청 형태에 대한 견적을 받으려면 `POST /api/v1/video/quote`를 호출하세요. quote endpoint가 유일한 권위 있는 출처입니다. 가격 세부 사항은 변경될 수 있으며 클라이언트 측에서 캐시하거나 복제해서는 안 됩니다.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+레퍼런스 비디오가 요청에 포함되어 있으면, 견적이 `/video/queue`가 부과하는 금액과 일치하도록 `reference_video_total_duration`(모든 레퍼런스 클립 길이의 합, 초 단위)도 전달하세요:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,9 +221,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## 완전한 예시
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+모든 예시는 `VENICE_API_KEY`가 환경에 설정되어 있다고 가정합니다.
 
 ### Text-to-video
 
@@ -239,7 +232,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### Image-to-video(first frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video`(와 그의 Fast 변형)는 **`aspect_ratio`를 받지 않습니다** — 출력 종횡비는 입력 이미지의 크기에서 자동 파생됩니다. 필드를 전달하면 *"This model does not support aspect_ratio"* 와 함께 400을 반환합니다. 명시적 종횡비 제어가 필요하면 T2V 또는 R2V 변형을 사용하세요.
 </Note>
 
-### Reference workflow — subject donor
+### Reference 워크플로 — 피사체 도너
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Reference 워크플로 — 피사체 + 오디오 도너
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Edit 워크플로
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### 이미지 그라운딩이 있는 Edit 워크플로
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -390,7 +334,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -400,14 +344,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clips)
+### Stitch(3개 클립)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### 완료 폴링
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+큐 제출 후 반환된 `queue_id`를 저장하고 응답 본문이 `video/mp4`가 될 때까지 `/video/retrieve`로 폴링하세요:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+응답은 작업이 완료될 때까지 JSON(`{ "status": "queued" | "running" | "failed", ... }`)이며, 완료되면 응답 본문이 `video/mp4` 바이트로 전환됩니다. 전체 폴링 패턴은 [Video Generation](/guides/media/video-generation)을 참고하세요.
 
 ---
 
-## Troubleshooting
+## 문제 해결
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Reference-to-video 제출에는 `reference_image_urls`, `reference_video_urls`, `image_references`, `video_references` 중 최소 하나가 포함되어야 합니다. 순수 텍스트만의 생성은 유효한 R2V 워크플로가 아닙니다 — 대신 `seedance-2-0-text-to-video`를 사용하세요. `reference_audio_urls`만으로는 충분하지 않습니다(위 Audio 섹션 참고).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+모델은 레퍼런스 비디오를 3개로 제한합니다. 더 많은 클립이 필요하면 먼저 Stitch를 실행(3 → 1)한 다음 그 출력을 후속 작업의 레퍼런스로 사용하세요.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / 합계 `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+클립당 길이는 `[2, 15]` 초 **포함**이며, 모든 레퍼런스 비디오의 합도 15초로 제한됩니다. 제출 전 클라이언트 측에서 클립을 트림하세요.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### Prompt가 잘못된 워크플로로 라우팅됨
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+워크플로는 prompt 문법에서 추론됩니다. 흔한 잘못된 라우팅:
 
-### Face / likeness inputs rejected
+- **Extend**를 원하지만 `Refer to ...`로 작성 → 모델이 비디오를 이어갈 *캔버스*가 아닌 *도너*로 취급
+- **Stitch**를 원하지만 `Refer to ...`로 작성 → 모델이 하나를 도너로 선택하고 나머지를 무시
+- **Edit**를 원하지만 `Generate a video based on <Video 1>`로 작성 → 모호함, 모델이 Reference로 기본 처리할 수 있음
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+표준 prefix를 정확히 작성된 대로 사용하세요: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### 견적이 큐 금액과 일치하지 않음
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+레퍼런스 비디오를 포함했지만 `reference_video_total_duration`을 `/video/quote`에 전달하지 않았다면 견적과 큐 금액이 다를 수 있습니다. 레퍼런스 비디오가 있을 때는 항상 `reference_video_total_duration`(모든 레퍼런스 클립 길이의 합, 초 단위)을 전달하세요.
 
 ---
 
-## References
+## 참고 자료
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Venice 비디오 큐 endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice 견적 endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- 동반 가이드: [Reference to Video](/guides/media/reference-to-video)(Kling O3 + Grok Imagine R2V 다룸)
+- 동반 가이드: [Video Generation](/guides/media/video-generation)(큐 / 폴링 개요)
diff --git a/ko/guides/media/video-generation.mdx b/ko/guides/media/video-generation.mdx
index 3e40c6e7..7bf1c6fb 100644
--- a/ko/guides/media/video-generation.mdx
+++ b/ko/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Image-to-video 모델의 경우 `image_url`을 통해 원본 이미지를 전달
 견적은 USD 단위입니다.
 
 <Info>
-**견적 파라미터는 가격 산정 input 전용입니다.** `/video/quote`에 전송하는 필드(`aspect_ratio`, `audio`, `reference_video_total_duration` 포함)는 순전히 가격을 계산하는 데만 사용됩니다. 이 필드들은 `/video/queue`로 전달되지 **않으며** 생성되는 비디오에 영향을 주지 않습니다. quote에 `aspect_ratio`를 보내는 것은 생성 시점에 `aspect_ratio`를 거부하는 모델(예: `seedance-2-0-image-to-video-basic`)에도 허용되는데, 이는 image-to-video 모델이 입력 이미지에서 출력 종횡비를 도출하기 때문입니다.
+**견적 파라미터는 가격 산정 input 전용입니다.** `/video/quote`에 전송하는 필드(`aspect_ratio`, `audio`, `reference_video_total_duration` 포함)는 순전히 가격을 계산하는 데만 사용됩니다. 이 필드들은 `/video/queue`로 전달되지 **않으며** 생성되는 비디오에 영향을 주지 않습니다. quote에 `aspect_ratio`를 보내는 것은 생성 시점에 `aspect_ratio`를 거부하는 모델(예: `seedance-2-0-image-to-video`)에도 허용되는데, 이는 image-to-video 모델이 입력 이미지에서 출력 종횡비를 도출하기 때문입니다.
 </Info>
 
 ### Reference-to-video 가격
diff --git a/pt-BR/api-reference/endpoint/video/queue.mdx b/pt-BR/api-reference/endpoint/video/queue.mdx
index dd58d2fa..9872764a 100644
--- a/pt-BR/api-reference/endpoint/video/queue.mdx
+++ b/pt-BR/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ Chame `/video/quote` para obter uma estimativa de preço e, em seguida, consulte
 
 Os modelos privados também retornam uma `download_url` para o vídeo finalizado. É uma URL de entrega de curta duração (algumas tentativas são aceitáveis se um download falhar); veja o [guia de geração de vídeo](/guides/media/video-generation#private-download-links) para detalhes e o `DELETE` opcional para privacidade.
 
-### Seedance 2.0 e 2.5
+### Seedance 2.0
 
-Para os modelos `seedance-2-0-*-basic` e `seedance-2-5-*-basic` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*` do Seedance 2.0), consulte o [Guia Seedance 2.0 e 2.5](/guides/media/seedance-2-0) para o modelo de quatro fluxos (Reference / Edit / Extend / Stitch), limites multimodais por família, comportamento faces-off da API e detalhes de preços.
+Para modelos `seedance-2-0-*` (text-to-video, image-to-video, reference-to-video, além das variantes `-fast-*`), consulte o [Guia do Seedance 2.0](/guides/media/seedance-2-0) para o modelo de quatro workflows (Reference / Edit / Extend / Stitch), padrões canônicos de prompt, limites de entrada multimodal e detalhes de preço.
 
-### API Seedance faces-off
+### Consentimento Seedance
 
-Os modelos Seedance públicos **não** suportam provisionamento de face assets nem atestação de consentimento facial. Entradas com rostos podem ser rejeitadas upstream. Para geração com likeness / face assets, use o app ou o Studio da Venice. Veja o [guia Seedance](/guides/media/seedance-2-0).
+Quando uma requisição image- ou reference-to-video do Seedance contém um rosto humano, a API retorna `409 needs_consent` com o texto da política, e você reenvia a mesma requisição com uma atestação `consents.seedance`. Veja o [Guia de consentimento para mídia facial do Seedance](/guides/media/seedance-face-consent) para o fluxo completo, o objeto de consentimento, deduplicação e revogação.
 
 ### Upscaling de vídeo
 
diff --git a/pt-BR/guides/integrations/venice-video-harness.mdx b/pt-BR/guides/integrations/venice-video-harness.mdx
index 7b2db68e..50eb08a8 100644
--- a/pt-BR/guides/integrations/venice-video-harness.mdx
+++ b/pt-BR/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ Os padrões do harness são opinativos porque consistência é o ponto. Roteamen
 
 | Papel | Modelo padrão | Quando usado |
 |---|---|---|
-| Tomadas de personagem (1-2 personagens) | `seedance-2-0-reference-to-video-basic` | R2V padrão com `reference_image_urls` flat, tags `@Image`, até 15s, áudio estéreo nativo |
+| Tomadas de personagem (1-2 personagens) | `seedance-2-0-reference-to-video` | R2V padrão com `reference_image_urls` flat, tags `@Image`, até 15s, áudio estéreo nativo |
 | Tomadas de personagem (3+ personagens) | `kling-o3-standard-reference-to-video` | Auto-fallback com `elements` estruturados para identidade multi-personagem |
-| Estabelecimento / atmosfera / ação | `seedance-2-0-image-to-video-basic` | Sem personagens; qualidade cinemática épica, até 15s |
+| Estabelecimento / atmosfera / ação | `seedance-2-0-image-to-video` | Sem personagens; qualidade cinemática épica, até 15s |
 
 Esses são sobrescritíveis por projeto via `series.json → videoDefaults`. Para mirar em uma família que não seja Seedance (por exemplo, contas que não têm acesso ao Seedance), defina `videoDefaults` como `kling-o3-standard-reference-to-video` e `veo3.1-fast-image-to-video`.
 
diff --git a/pt-BR/guides/media/seedance-2-0.mdx b/pt-BR/guides/media/seedance-2-0.mdx
index be4cbc87..e96bfa21 100644
--- a/pt-BR/guides/media/seedance-2-0.mdx
+++ b/pt-BR/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 e 2.5"
-description: "Gere, edite, estenda e una vídeos com Seedance 2.0 e 2.5 na Venice - fluxos texto/imagem/referência, API faces-off e limites multimodais por família."
-'og:title': "Guia Seedance 2.0 e 2.5 | Venice API Docs"
-'og:description': "Gere, edite, estenda e una vídeos com Seedance 2.0 e 2.5 via API Venice. Fluxos, faces-off, limites, preços e exemplos."
+title: "Seedance 2.0"
+description: "Gere, edite, estenda e costure vídeos com o Seedance 2.0 na Venice — fluxos de texto, imagem e reference-to-video em um único endpoint, com exemplos práticos."
+'og:title': "Guia do Seedance 2.0 | Documentação da API Venice"
+'og:description': "Gere, edite, estenda e costure vídeos com o Seedance 2.0 via a API Venice. Padrões de workflow, limites multimodais, preços e exemplos completos."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+O Seedance 2.0 é um modelo multimodal de vídeo flagship exposto na Venice como uma família de três variantes para geração de vídeo guiada por texto, imagem e referência. A variante **reference-to-video** é excepcionalmente poderosa: um único endpoint e um único ID de modelo cuidam de **quatro workflows distintos** (Reference, Edit, Extend, Stitch) — o workflow é inferido pelo **formato do seu prompt**.
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+Este guia percorre as variantes, os quatro workflows com seus prompts canônicos, os limites de entrada multimodais, preços e exemplos completos em `curl`.
 
-<Warning>
-**Rostos não são suportados na API pública do Seedance.** A Venice não executa provisionamento de face assets nem atestação de consentimento facial (`consents.seedance` / `needs_consent`) para modelos Seedance públicos. Entradas com rostos podem ser rejeitadas upstream como erro de política de conteúdo / provedor. Para geração com likeness / face assets, use o app ou o Studio da Venice.
-</Warning>
-
-## Variants
+## Variantes
 
-| Model ID | Variant | Output resolutions | Notes |
+| ID do modelo | Variante | Resoluções de saída | Notas |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | Apenas prompt de texto |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | Grounding por imagem do primeiro frame (e opcionalmente último frame) |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | Até 9 imagens de referência + 3 vídeos de referência + 3 áudios de referência. Alimenta Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | Tier mais rápido e de menor fidelidade |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | Tier mais rápido e de menor fidelidade |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | Tier mais rápido e de menor fidelidade; mesmo conjunto de workflows |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+Todas as variantes são assíncronas. Envie via `POST /api/v1/video/queue`, e depois faça polling em `POST /api/v1/video/retrieve` até o corpo da resposta ser `video/mp4`. Veja [Geração de vídeo](/guides/media/video-generation) para o fluxo geral de fila.
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## O modelo "um modelo, quatro workflows"
 
-## The "one model, four workflows" model
+A variante reference-to-video (`seedance-2-0-reference-to-video` e sua versão Fast) é o mesmo modelo subjacente servindo quatro tarefas diferentes. **O modelo infere a tarefa a partir do prefixo do prompt e do formato das suas entradas.** Não há campo `task` ou `workflow` — a sintaxe do prompt é o roteamento.
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| Workflow | O que faz | Prefixo do prompt | Entradas |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | Gera um novo vídeo usando arquivos de referência enviados como doadores para sujeito / movimento / estilo / áudio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Texto + ≥1 imagem OU vídeo de referência (0-9 imagens, 0-3 vídeos), mais opcionalmente até 3 doadores de áudio |
+| **Edit** | Modifica um único vídeo de entrada preservando o resto | `Strictly edit <Video 1>, changing its ...` | 1 vídeo de entrada + texto (imagens são grounding opcional) |
+| **Extend** | Extensão para frente / para trás de um clipe | `Extend <Video 1>, generate ...` | 1 vídeo de entrada + texto |
+| **Stitch** | Costura 2-3 clipes com transições geradas automaticamente | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 vídeos de entrada + texto |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+A **sintaxe do prompt é canônica e sensível a maiúsculas**: colchetes angulares, primeira letra maiúscula, um único espaço antes do número — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
 
 ---
 
-## Workflow patterns
+## Padrões de workflow
 
-### Reference workflow
+### Workflow Reference
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+Use os arquivos de referência enviados como **doadores** — sujeito, cena, movimento, estilo, timbre vocal — para gerar um vídeo totalmente novo.
 
-**Canonical prompt patterns**:
+**Padrões de prompt canônicos**:
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,23 +51,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**Exemplos**:
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (doadores de áudio devem ser pareados com pelo menos uma referência de imagem ou vídeo — apenas áudio é rejeitado)
 
-### Edit workflow
+### Workflow Edit
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+Modifique um único vídeo de entrada. **Qualquer coisa que não seja explicitamente nomeada no prompt é preservada.** Use isso quando você quer uma mudança localizada (troca de sujeito, mudança de clima/cor, adição/remoção de elemento) em vez de um vídeo totalmente novo.
 
-**Canonical prompt pattern**:
+**Padrão de prompt canônico**:
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**Subpadrões para controle mais fino**:
 
 ```
 Add Elements:
@@ -89,57 +80,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**Exemplos**:
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+O último exemplo combina Edit com uma referência de imagem — perfeitamente legal, o modelo usa `<Image 1>` como doador visual para a substituição.
 
-### Extend workflow
+### Workflow Extend
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+Continue um único clipe para frente ou para trás no tempo. **Por padrão, o Seedance retorna apenas o novo conteúdo** — não o input original concatenado com a extensão. Isso é por design, para continuidade da transição; se você quiser que o clipe de input seja preservado junto com a extensão, diga isso explicitamente:
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserva input no início
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserva input no fim
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+Tratamento de transição: o modelo extrai automaticamente os frames de transição para mesclagem suave, e os segmentos originais do vídeo de input não são regerados.
 
-**Examples**:
+**Exemplos**:
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Workflow Stitch (Track Completion)
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+Conecte 2-3 clipes de input com transições geradas por IA. A duração total combinada de input deve ser ≤ 15 s.
 
-**Canonical prompt pattern**:
+**Padrão de prompt canônico**:
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**Exemplos**:
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+O modelo apara automaticamente os segmentos de conexão nos pontos de junção para continuidade.
 
 ---
 
-## Universal prompt formula
+## Fórmula de prompt universal
 
-Across all four workflows, the recommended authoring formula is:
+Em todos os quatro workflows, a fórmula de autoria recomendada é:
 
 ```
 Subject + Motion + Environment (Optional)
@@ -148,77 +139,79 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subject + Motion**: a base lógica — define "Quem" está realizando "Qual ação"
+- **Environment + Aesthetics**: plano de fundo espacial, iluminação, estilo visual
+- **Camera**: tipo de plano ou movimento explícito
+- **Audio**: efeitos sonoros ambientes ou direção vocal para saída imersiva
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+Sobrepor isso ao prefixo de um workflow (por exemplo, `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produz as saídas de mais alta qualidade.
 
 ---
 
-## Multimodal input limits
+## Limites de entrada multimodal
+
+Os valores abaixo são o que a API Venice aceita. Requisições fora dessas faixas são rejeitadas na camada de schema com um 400 antes de chegar à inferência.
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### Imagens
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| Restrição | Valor |
+|---|---|
+| Métodos de input | URL (`http://`, `https://`) ou Base64 data URL (`data:image/...`) |
+| Formatos | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
+| Aspect ratio (W / H) | exclusivo `(0.4, 2.5)` |
+| Lado mínimo | ≥ 300 px |
+| Quantidade de imagem: primeiro frame I2V | 1 |
+| Quantidade de imagem: I2V primeiro + último frame | 2 |
+| **Quantidade de imagem: R2V (V2 / Fast)** | **1 – 9** |
 
-### Shared media floors
+### Vídeos
 
-| Constraint | Value |
+| Restrição | Valor |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| Métodos de input | URL (`http://`, `https://`) ou Base64 data URL (`data:video/...`) |
+| Formatos | `.mp4`, `.mov` |
+| Codecs de vídeo | H.264 / AVC, H.265 / HEVC |
+| Codecs de áudio (no container) | AAC, MP3 |
+| **Duração por clipe** | `[2, 15]` s (inclusivo) |
+| **Máximo de clipes** | 3 (R2V / Stitch / Extend) |
+| **Duração total combinada** | ≤ 15 s em todos os clipes |
+| **Tamanho por clipe** | ≤ 50 MB |
+
+### Áudio
+
+| Restrição | Valor |
+|---|---|
+| Métodos de input | URL (`http://`, `https://`) ou Base64 data URL (`data:audio/...`) |
+| Formatos | `.wav`, `.mp3` |
+| **Duração por clipe** | `[2, 15]` s |
+| **Máximo de clipes** | 3 |
+| **Duração total combinada** | ≤ 15 s em todos os clipes |
+| **Tamanho por clipe** | ≤ 15 MB |
+
+Áudio de referência é suportado apenas nas variantes R2V. Cada entrada é encaminhada ao modelo como um item de conteúdo `role: "reference_audio"` que o prompt endereça como `<Audio 1>`, `<Audio 2>`, `<Audio 3>` — o modelo usa cada clipe para timbre vocal, efeitos sonoros ou música de fundo dependendo de como o prompt o enquadra. O campo legado singular `audio_url` mapeia para o mesmo formato de conteúdo e agora é equivalente a passar um `reference_audio_urls` com um único elemento.
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls` não pode ser a única entrada de referência.** O modelo requer pelo menos uma referência de imagem ou vídeo junto com qualquer doador de áudio. Pareie `reference_audio_urls` com `reference_image_urls`, `reference_video_urls`, `image_url` ou `video_url` — envios apenas de áudio são rejeitados.
 </Warning>
 
-### Request size
+### Tamanho da requisição
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+O endpoint de queue aceita corpos JSON de até **35 MB**. Data URLs inline para vídeos grandes podem ultrapassar isso — para Stitch multi-clipe em particular, prefira URLs em vez de base64 inline.
 
 ---
 
-## Pricing
+## Preços
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+Chame `POST /api/v1/video/quote` para obter uma cotação para um determinado formato de requisição antes de enviá-la a `/video/queue`. O endpoint de cotação é a única fonte oficial; detalhes de preços podem mudar e não devem ser cacheados ou duplicados no cliente.
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+Quando vídeo(s) de referência fizerem parte da requisição, também passe `reference_video_total_duration` (a soma de todas as durações de clipes de referência em segundos) para que a cotação corresponda ao que `/video/queue` vai cobrar:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,9 +221,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## Exemplos completos
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+Todos os exemplos assumem que `VENICE_API_KEY` está definido no ambiente.
 
 ### Text-to-video
 
@@ -239,7 +232,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### Image-to-video (primeiro frame)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video` (e sua variante Fast) **não aceitam `aspect_ratio`** — o aspect ratio de saída é auto-derivado das dimensões da imagem de input. Passar o campo retorna um 400 com *"This model does not support aspect_ratio"*. Use as variantes T2V ou R2V se precisar de controle explícito de aspect ratio.
 </Note>
 
-### Reference workflow — subject donor
+### Workflow Reference — doador de sujeito
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Workflow Reference — doador de sujeito + áudio
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Workflow Edit
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### Workflow Edit com grounding de imagem
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### Extend para frente
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -400,14 +344,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clips)
+### Stitch (3 clipes)
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### Polling para conclusão
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+Após cada envio à fila, salve o `queue_id` retornado e faça polling em `/video/retrieve` até o corpo da resposta ser `video/mp4`:
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+A resposta é JSON (`{ "status": "queued" | "running" | "failed", ... }`) até o job ser concluído, momento em que o corpo da resposta passa para bytes `video/mp4`. Veja [Geração de vídeo](/guides/media/video-generation) para o padrão completo de polling.
 
 ---
 
-## Troubleshooting
+## Solução de problemas
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Envios de reference-to-video devem incluir pelo menos um de `reference_image_urls`, `reference_video_urls`, `image_references` ou `video_references`. Geração apenas com texto não é um workflow R2V válido — use `seedance-2-0-text-to-video` em vez disso. `reference_audio_urls` sozinho não é suficiente (veja a seção Áudio acima).
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+O modelo limita os vídeos de referência a 3. Se precisar de mais clipes, execute um Stitch primeiro (3 → 1) e depois use a saída como referência para uma sequência.
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / agregado `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+Duração por clipe é `[2, 15]` segundos **inclusivo**; a soma em todos os vídeos de referência também é limitada a 15 segundos. Apare os clipes no cliente antes do envio.
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### O prompt vai para o workflow errado
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+O workflow é inferido pela sintaxe do prompt. Erros comuns de roteamento:
 
-### Face / likeness inputs rejected
+- Querer **Extend** mas escrever `Refer to ...` → o modelo trata seu vídeo como um *doador*, não como um canvas a continuar
+- Querer **Stitch** mas escrever `Refer to ...` → o modelo escolhe um como doador, ignora os outros
+- Querer **Edit** mas escrever `Generate a video based on <Video 1>` → ambíguo; o modelo pode optar por Reference
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+Use os prefixos canônicos exatamente como escritos: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
 
-### Quote doesn't match the queued amount
+### A cotação não corresponde ao valor enfileirado
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+Se você incluiu um vídeo de referência mas não passou `reference_video_total_duration` para `/video/quote`, a cotação e o valor enfileirado podem diferir. Sempre passe `reference_video_total_duration` (soma de todas as durações dos clipes de referência, em segundos) quando vídeos de referência estiverem presentes.
 
 ---
 
-## References
+## Referências
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Endpoint de fila de vídeo da Venice: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Endpoint de cotação da Venice: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- Guia companheiro: [Reference to Video](/guides/media/reference-to-video) (cobre Kling O3 + Grok Imagine R2V)
+- Guia companheiro: [Geração de vídeo](/guides/media/video-generation) (visão geral de fila / polling)
diff --git a/pt-BR/guides/media/video-generation.mdx b/pt-BR/guides/media/video-generation.mdx
index a7c944e4..b98acfdb 100644
--- a/pt-BR/guides/media/video-generation.mdx
+++ b/pt-BR/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ Obtenha o custo exato antes de gerar. Envie apenas inputs de preço (`model`, `d
 A cotação está em USD.
 
 <Info>
-**Os parâmetros de cotação são apenas inputs de preço.** Os campos que você envia para `/video/quote` — incluindo `aspect_ratio`, `audio` e `reference_video_total_duration` — são usados exclusivamente para calcular o preço. Eles **não** são encaminhados para `/video/queue` e não afetam o vídeo gerado. Enviar `aspect_ratio` para a cotação é aceito mesmo para modelos (como `seedance-2-0-image-to-video-basic`) que rejeitam `aspect_ratio` no momento da geração, já que modelos image-to-video derivam o aspect ratio de saída a partir da imagem de entrada.
+**Os parâmetros de cotação são apenas inputs de preço.** Os campos que você envia para `/video/quote` — incluindo `aspect_ratio`, `audio` e `reference_video_total_duration` — são usados exclusivamente para calcular o preço. Eles **não** são encaminhados para `/video/queue` e não afetam o vídeo gerado. Enviar `aspect_ratio` para a cotação é aceito mesmo para modelos (como `seedance-2-0-image-to-video`) que rejeitam `aspect_ratio` no momento da geração, já que modelos image-to-video derivam o aspect ratio de saída a partir da imagem de entrada.
 </Info>
 
 ### Preços de reference-to-video
diff --git a/zh/api-reference/endpoint/video/queue.mdx b/zh/api-reference/endpoint/video/queue.mdx
index 5b01b9e0..e1852398 100644
--- a/zh/api-reference/endpoint/video/queue.mdx
+++ b/zh/api-reference/endpoint/video/queue.mdx
@@ -9,13 +9,13 @@ openapi: 'POST /video/queue'
 
 私有模型还会为已完成的视频返回 `download_url`。这是一个短期分发 URL（下载失败时几次重试是可以的）；详情和可选的 `DELETE` 隐私操作请参阅[视频生成指南](/guides/media/video-generation#private-download-links)。
 
-### Seedance 2.0 与 2.5
+### Seedance 2.0
 
-对于 `seedance-2-0-*-basic` 与 `seedance-2-5-*-basic` 模型（text-to-video、image-to-video、reference-to-video，以及 Seedance 2.0 的 `-fast-*` 变体），请参阅 [Seedance 2.0 与 2.5 指南](/guides/media/seedance-2-0)，了解四工作流模型（Reference / Edit / Extend / Stitch）、按系列区分的多模态限制、faces-off API 行为与定价说明。
+对于 `seedance-2-0-*` 模型（文生视频、图生视频、参考生视频以及 `-fast-*` 变体），请参阅 [Seedance 2.0 指南](/guides/media/seedance-2-0) 了解四工作流模型（Reference / Edit / Extend / Stitch）、规范的 prompt 模式、多模态输入限制和定价详情。
 
-### Seedance faces-off API
+### Seedance 同意声明
 
-公共 Seedance 模型**不**支持人脸资产开通或人脸同意证明。含人脸的输入可能会在上游被拒绝。如需 likeness / 人脸资产生成，请使用 Venice 应用或 Studio。详见 [Seedance 指南](/guides/media/seedance-2-0)。
+当 Seedance 的图生视频或参考生视频请求包含人脸时，API 会返回带有政策文本的 `409 needs_consent`，您需要在重新提交相同请求时附带 `consents.seedance` 证明。完整流程、同意对象、去重和撤销机制请参阅 [Seedance 人脸媒体同意指南](/guides/media/seedance-face-consent)。
 
 ### 视频放大
 
diff --git a/zh/guides/integrations/venice-video-harness.mdx b/zh/guides/integrations/venice-video-harness.mdx
index 877280e0..d72aaf9d 100644
--- a/zh/guides/integrations/venice-video-harness.mdx
+++ b/zh/guides/integrations/venice-video-harness.mdx
@@ -124,9 +124,9 @@ harness 默认值是有主见的，因为一致性是关键。当前路由（202
 
 | 角色 | 默认模型 | 使用场景 |
 |---|---|---|
-| 角色镜头（1-2 个角色） | `seedance-2-0-reference-to-video-basic` | 默认 R2V，采用扁平 `reference_image_urls`、`@Image` 标签，最长 15 秒，原生立体声音频 |
+| 角色镜头（1-2 个角色） | `seedance-2-0-reference-to-video` | 默认 R2V，采用扁平 `reference_image_urls`、`@Image` 标签，最长 15 秒，原生立体声音频 |
 | 角色镜头（3+ 个角色） | `kling-o3-standard-reference-to-video` | 自动回退，采用结构化 `elements` 处理多角色身份 |
-| 建立镜头/氛围/动作 | `seedance-2-0-image-to-video-basic` | 无角色；史诗级电影质量，最长 15 秒 |
+| 建立镜头/氛围/动作 | `seedance-2-0-image-to-video` | 无角色；史诗级电影质量，最长 15 秒 |
 
 这些可以通过 `series.json → videoDefaults` 按项目覆盖。要针对非 Seedance 系列（例如缺乏 Seedance 访问权限的账户），将 `videoDefaults` 设置为 `kling-o3-standard-reference-to-video` 和 `veo3.1-fast-image-to-video`。
 
diff --git a/zh/guides/media/seedance-2-0.mdx b/zh/guides/media/seedance-2-0.mdx
index 680289fc..88dee047 100644
--- a/zh/guides/media/seedance-2-0.mdx
+++ b/zh/guides/media/seedance-2-0.mdx
@@ -1,58 +1,49 @@
 ---
-title: "Seedance 2.0 与 2.5"
-description: "使用 Venice 上的 Seedance 2.0 与 2.5 生成、编辑、延长并拼接视频 - 文本/图像/参考工作流、faces-off API 行为，以及按系列区分的多模态限制。"
-'og:title': "Seedance 2.0 与 2.5 指南 | Venice API Docs"
-'og:description': "通过 Venice API 使用 Seedance 2.0 与 2.5 生成、编辑、延长并拼接视频。工作流、faces-off、多模态限制、定价与示例。"
+title: "Seedance 2.0"
+description: "在 Venice 上使用 Seedance 2.0 生成、编辑、延展并拼接视频 —— 在同一个端点上提供文本、图像和 reference-to-video 工作流，并附带实战示例。"
+'og:title': "Seedance 2.0 Guide | Venice API Docs"
+'og:description': "Generate, edit, extend and stitch videos with Seedance 2.0 via the Venice API. Workflow patterns, multimodal limits, pricing, and worked examples."
 ---
 
-Seedance is a flagship multimodal video family on Venice for text-, image-, and reference-driven video generation. **Seedance 2.0** (plus Fast) and **Seedance 2.5** share the same R2V prompt-routing model: a single reference-to-video endpoint handles **four distinct workflows** (Reference, Edit, Extend, Stitch) — the workflow is inferred from the **shape of your prompt**.
+Seedance 2.0 是一个旗舰多模态视频模型，Venice 将其暴露为三个变体的系列，用于文本驱动、图像驱动和参考驱动的视频生成。**reference-to-video** 变体异常强大：单个端点和单个模型 ID 处理**四种不同的工作流**（Reference、Edit、Extend、Stitch）——工作流从您的 **prompt 形态**推断而出。
 
-This guide covers variants, the four workflows, **faces-off public API behavior**, **family-specific multimodal limits**, pricing, and `curl` examples.
+本指南介绍各变体、四种工作流及其规范 prompt、多模态输入限制、定价和完整的 `curl` 示例。
 
-<Warning>
-**公共 Seedance API 不支持人脸。** Venice 不会为公共 Seedance 模型运行人脸资产开通或人脸同意证明（`consents.seedance` / `needs_consent`）。含人脸的输入可能在上游被以内容政策 / 提供商错误拒绝。如需 likeness / 人脸资产生成，请使用 Venice 应用或 Studio。
-</Warning>
-
-## Variants
+## 变体
 
-| Model ID | Variant | Output resolutions | Notes |
+| 模型 ID | 变体 | 输出分辨率 | 备注 |
 |---|---|---|---|
-| `seedance-2-0-text-to-video-basic` | T2V | 480p / 720p / 1080p / **4k** | Text prompt only |
-| `seedance-2-0-image-to-video-basic` | I2V | 480p / 720p / 1080p / **4k** | First-frame (and optionally last-frame) image grounding |
-| `seedance-2-0-reference-to-video-basic` | R2V | 480p / 720p / 1080p / **4k** | Up to 9 reference images + 3 reference videos + 3 reference audio donors. Powers Reference / Edit / Extend / Stitch |
-| `seedance-2-0-fast-text-to-video-basic` | Fast T2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-image-to-video-basic` | Fast I2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k) |
-| `seedance-2-0-fast-reference-to-video-basic` | Fast R2V | 480p / 720p | Faster, lower-fidelity tier (no 1080p / 4k); same workflow set |
-| `seedance-2-5-text-to-video-basic` | T2V | 480p / 720p | Up to 30s output; native audio |
-| `seedance-2-5-image-to-video-basic` | I2V | 480p / 720p | Up to 30s output; first-frame (and optionally last-frame) grounding |
-| `seedance-2-5-reference-to-video-basic` | R2V | 480p / 720p | Up to 30 images + 10 videos + 10 audio donors; same Reference / Edit / Extend / Stitch workflows |
+| `seedance-2-0-text-to-video` | T2V | 480p / 720p / 1080p | 仅文本 prompt |
+| `seedance-2-0-image-to-video` | I2V | 480p / 720p / 1080p | 第一帧（及可选最后一帧）图像基础 |
+| `seedance-2-0-reference-to-video` | R2V | 480p / 720p / 1080p | 最多 9 张参考图像 + 3 段参考视频 + 3 段参考音频供体。驱动 Reference / Edit / Extend / Stitch |
+| `seedance-2-0-fast-text-to-video` | Fast T2V | 480p / 720p | 更快、较低保真度层级 |
+| `seedance-2-0-fast-image-to-video` | Fast I2V | 480p / 720p | 更快、较低保真度层级 |
+| `seedance-2-0-fast-reference-to-video` | Fast R2V | 480p / 720p | 更快、较低保真度层级；相同工作流集 |
 
-All variants are async. Submit via `POST /api/v1/video/queue`, then poll `POST /api/v1/video/retrieve` until the response body is `video/mp4`. See [Video Generation](/guides/media/video-generation) for the general queue flow.
+所有变体都是异步的。通过 `POST /api/v1/video/queue` 提交，然后轮询 `POST /api/v1/video/retrieve` 直到响应体为 `video/mp4`。常规队列流程请参阅[视频生成](/guides/media/video-generation)。
 
-Pass `resolution` as one of: `480p`, `720p`, `1080p`, or `4k` (lowercase). Seedance **2.0** (non-Fast) accepts all four; **2.0 Fast** and **2.5** accept `480p` / `720p` only. Discover live model IDs and capabilities with `GET /models?type=video` — do not hardcode availability.
+## "一个模型，四个工作流"模型
 
-## The "one model, four workflows" model
+reference-to-video 变体（`seedance-2-0-reference-to-video` 及其 Fast 兄弟）是同一个底层模型服务四种不同任务。**模型从 prompt 前缀和您输入的形态推断任务。** 没有 `task` 或 `workflow` 字段——prompt 语法即路由。
 
-The reference-to-video variants (`seedance-2-0-reference-to-video-basic`, its Fast sibling, and `seedance-2-5-reference-to-video-basic`) use the same prompt-routing pattern. **The model infers the task from the prompt prefix and the shape of your inputs.** There is no `task` or `workflow` field — the prompt syntax is the routing.
-
-| Workflow | What it does | Prompt prefix | Inputs |
+| 工作流 | 功能 | Prompt 前缀 | 输入 |
 |---|---|---|---|
-| **Reference** | Generate a new video using uploaded reference files as donors for subject / motion / style / audio | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | Text + ≥1 image OR video reference, plus optional audio donors (counts differ by family — see limits) |
-| **Edit** | Modify a single input video while preserving the rest | `Strictly edit <Video 1>, changing its ...` | 1 input video + text (images optional grounding) |
-| **Extend** | Forward / backward extension of one clip | `Extend <Video 1>, generate ...` | 1 input video + text |
-| **Stitch** | Stitch input clips with auto-generated transitions | `<Video 1> + <transition description> + followed by <Video 2> + ...` | Multiple input videos + text (clip count / duration caps differ by family) |
+| **Reference** | 使用上传的参考文件作为主体/动作/风格/音频的供体生成新视频 | `Refer to ... in <Image\|Video\|Audio N> to generate ...` | 文本 + ≥1 张图像 或 视频参考（0-9 张图像，0-3 段视频），可选最多 3 段音频供体 |
+| **Edit** | 修改单个输入视频同时保留其余部分 | `Strictly edit <Video 1>, changing its ...` | 1 段输入视频 + 文本（图像可选基础） |
+| **Extend** | 单个剪辑的前向/后向扩展 | `Extend <Video 1>, generate ...` | 1 段输入视频 + 文本 |
+| **Stitch** | 用自动生成的过渡拼接 2-3 段剪辑 | `<Video 1> + <transition description> + followed by <Video 2> + ...` | 2-3 段输入视频 + 文本 |
 
-The **prompt syntax is canonical and case-sensitive**: angle brackets, capital first letter, single space before the number — `<Video 1>`, `<Image 1>`, `<Audio 1>`.
+**prompt 语法是规范的且区分大小写**：尖括号、首字母大写、数字前单个空格——`<Video 1>`、`<Image 1>`、`<Audio 1>`。
 
 ---
 
-## Workflow patterns
+## 工作流模式
 
-### Reference workflow
+### Reference 工作流
 
-Use the uploaded reference files as **donors** — subject, scene, motion, style, vocal timbre — to generate a brand-new video.
+将上传的参考文件用作**供体**——主体、场景、动作、风格、人声音色——以生成全新的视频。
 
-**Canonical prompt patterns**:
+**规范 prompt 模式**：
 
 ```
 Refer to <Subject N> in <Image N> to generate ...
@@ -60,23 +51,23 @@ Refer to the [action | camera scene | style | sound effect] in <Video N> to gene
 Refer to the [tone | timbre] in <Audio N> to generate ...
 ```
 
-**Examples**:
+**示例**：
 
 - `Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character riding a horse through snow.`
 - `Refer to the camera scene in <Video 1> to generate a similar establishing shot of a futuristic city at dawn.`
-- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.` (audio donors must be paired with at least one image or video reference — audio alone is rejected)
+- `Refer to <Subject 1> in <Image 1> and use the timbre in <Audio 1> for the narrator describing the scene.`（音频供体必须与至少一张图像或视频参考配对——仅音频会被拒绝）
 
-### Edit workflow
+### Edit 工作流
 
-Modify a single input video. **Anything not explicitly named in the prompt is preserved.** Use this when you want a localized change (subject swap, weather/color change, element add/remove) rather than a wholly new video.
+修改单个输入视频。**prompt 中未明确命名的任何内容都会被保留。** 当您想要局部更改（主体替换、天气/颜色变化、元素添加/移除）而非全新视频时使用此项。
 
-**Canonical prompt pattern**:
+**规范 prompt 模式**：
 
 ```
 Strictly edit <Video 1>, changing its [original feature] to [new feature] ...
 ```
 
-**Sub-patterns for finer control**:
+**更精细控制的子模式**：
 
 ```
 Add Elements:
@@ -89,57 +80,57 @@ Modify Elements:
   Replace [description of element to be changed] in <Video 1> with [description of intended element].
 ```
 
-**Examples**:
+**示例**：
 
 - `Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm.`
 - `Add snacks such as fried chicken and pizza to the countertop in <Video 1>.`
 - `Remove the red car from <Video 1>, keeping the rest of the video content unchanged.`
 - `Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.`
 
-The last example combines Edit with an image reference — perfectly legal, the model uses `<Image 1>` as a visual donor for the replacement.
+最后一个示例结合了 Edit 和图像参考——完全合法，模型使用 `<Image 1>` 作为替换的视觉供体。
 
-### Extend workflow
+### Extend 工作流
 
-Continue a single clip forward or backward in time. **By default Seedance returns only the new content** — not the original input concatenated with the extension. This is by design, for transition continuity; if you want the input clip preserved alongside the extension, say so explicitly:
+将单个剪辑在时间上向前或向后延续。**默认情况下 Seedance 仅返回新内容**——而不是与扩展连接的原始输入。这是出于设计目的，用于过渡连续性；如果您希望保留输入剪辑与扩展一起，请明确说明：
 
 ```
 Extend <Video 1>, generate [description of extended content]
 Extend <Video 1> backward, [description of extended content]
-Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← preserves input at start
-Extend <Video 1> backward, [description], and then end with <Video 1>               ← preserves input at end
+Extend <Video 1>, start with <Video 1>, then [description of extended content]      ← 在开头保留输入
+Extend <Video 1> backward, [description], and then end with <Video 1>               ← 在结尾保留输入
 ```
 
-Transition handling: the model automatically extracts the transition frames for seamless blending, and the original segments of the input video are not re-generated.
+过渡处理：模型自动提取过渡帧进行无缝混合，输入视频的原始片段不会被重新生成。
 
-**Examples**:
+**示例**：
 
 - `Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk.`
 - `Extend <Video 1> backward, the same character walking toward the camera before the original shot begins.`
 - `Extend <Video 1>, start with <Video 1>, then the camera pulls back to reveal a vast landscape.`
 
-### Stitch workflow (Track Completion)
+### Stitch 工作流（Track Completion）
 
-Connect input clips with AI-generated transitions. Respect the **family-specific** combined-duration and clip-count caps in [Multimodal input limits](#multimodal-input-limits) (Seedance 2.0: ≤3 clips / ≤15 s combined; Seedance 2.5: higher video caps).
+用 AI 生成的过渡连接 2-3 段输入剪辑。总组合输入时长必须 ≤ 15 s。
 
-**Canonical prompt pattern**:
+**规范 prompt 模式**：
 
 ```
 <Video 1> + [transition description] + followed by <Video 2> [+ [transition description] + followed by <Video 3>]
 ```
 
-**Examples**:
+**示例**：
 
 - `<Video 1> + a smooth seamless cut + followed by <Video 2>`
 - `<Video 1>. The moment a leaf falls to the ground, it sets off a special effect of golden particles. A gust of wind blows by, leading into <Video 2>.`
 - `<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>`
 
-The model auto-trims connecting segments at the join points for continuity.
+模型在连接点自动修剪连接片段以保持连续性。
 
 ---
 
-## Universal prompt formula
+## 通用 prompt 公式
 
-Across all four workflows, the recommended authoring formula is:
+在所有四种工作流中，推荐的撰写公式为：
 
 ```
 Subject + Motion + Environment (Optional)
@@ -148,77 +139,79 @@ Subject + Motion + Environment (Optional)
        + Audio (Optional)
 ```
 
-- **Subject + Motion**: the logical foundation — define "Who" is performing "What action"
-- **Environment + Aesthetics**: spatial background, lighting, visual style
-- **Camera**: explicit shot type or movement
-- **Audio**: ambient sound effects or vocal direction for immersive output
+- **Subject + Motion**：逻辑基础——定义"谁"在执行"什么动作"
+- **Environment + Aesthetics**：空间背景、光照、视觉风格
+- **Camera**：明确的镜头类型或运动
+- **Audio**：用于沉浸式输出的环境音效或人声方向
 
-Layering this on top of a workflow prefix (e.g., `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`) produces the highest-quality outputs.
+将其叠加在工作流前缀之上（例如 `Strictly edit <Video 1>, changing its <subject + motion + environment + ...>`）能产生最高质量的输出。
 
 ---
 
-## Multimodal input limits
+## 多模态输入限制
+
+下面的值是 Venice API 接受的内容。超出这些范围的请求在到达推理之前会在 schema 层以 400 拒绝。
 
-Values below are what the Venice API accepts. Requests outside these ranges are rejected at the schema layer with a 400 before reaching inference.
+### 图像
 
-**Seedance 2.0 and Seedance 2.5 use different caps.** Always check the column for the model family you are calling.
+| 约束 | 值 |
+|---|---|
+| 输入方法 | URL（`http://`、`https://`）或 Base64 data URL（`data:image/...`） |
+| 格式 | `.jpeg`、`.png`、`.webp`、`.bmp`、`.tiff`、`.gif`、`.heic`、`.heif` |
+| 宽高比（W / H） | 开区间 `(0.4, 2.5)` |
+| 最小边 | ≥ 300 px |
+| 图像数量：I2V 第一帧 | 1 |
+| 图像数量：I2V 第一 + 最后一帧 | 2 |
+| **图像数量：R2V（V2 / Fast）** | **1 – 9** |
 
-### Shared media floors
+### 视频
 
-| Constraint | Value |
+| 约束 | 值 |
 |---|---|
-| Image / video / audio input methods | URL (`http://`, `https://`) or Base64 data URL |
-| Image formats | `.jpeg`, `.png`, `.webp`, `.bmp`, `.tiff`, `.gif`, `.heic`, `.heif` |
-| Image aspect ratio (W / H) | exclusive `(0.4, 2.5)` |
-| Image minimum side | ≥ 300 px |
-| Video formats | `.mp4`, `.mov` |
-| Video codecs | H.264 / AVC, H.265 / HEVC |
-| Audio codecs (in container) | AAC, MP3 |
-| Audio formats (reference audio) | `.wav`, `.mp3` |
-| Per-clip video size | ≤ 50 MB |
-| Per-clip audio size | ≤ 15 MB |
-| I2V first-frame images | 1 |
-| I2V first + last frame | 2 |
-
-### Family comparison
-
-| Constraint | Seedance 2.0 (+ Fast) | Seedance 2.5 |
-|---|---|---|
-| Output duration | 4–15 s | 4–30 s (default 10) |
-| Output resolutions | 480p / 720p / 1080p / **4k** (Fast: 480p / 720p only) | 480p / 720p only |
-| R2V reference images | 1–9 | 1–30 |
-| Max R2V reference image bytes | (shared request limits) | ≤ 30 MB per image |
-| R2V reference videos | ≤ 3 | ≤ 10 |
-| Duration per reference video | `[2, 15]` s | `[2, 30]` s |
-| Combined reference video duration | ≤ 15 s | ≤ 30 s |
-| R2V reference audio clips | ≤ 3 | ≤ 10 |
-| Duration per reference audio | `[2, 15]` s | `[2, 30]` s |
-| Combined reference audio duration | ≤ 15 s | ≤ 30 s |
-
-Reference audio is supported on the R2V variants only. Each entry is forwarded to the model as a `role: "reference_audio"` content item that the prompt addresses as `<Audio 1>`, `<Audio 2>`, … — the model uses each clip for vocal timbre, sound effects, or background music depending on how the prompt frames it. The legacy singular `audio_url` field maps to the same content shape and is now equivalent to passing a one-element `reference_audio_urls`.
+| 输入方法 | URL（`http://`、`https://`）或 Base64 data URL（`data:video/...`） |
+| 格式 | `.mp4`、`.mov` |
+| 视频编解码器 | H.264 / AVC、H.265 / HEVC |
+| 音频编解码器（容器内） | AAC、MP3 |
+| **每个剪辑时长** | `[2, 15]` s（含端点） |
+| **最大剪辑数量** | 3（R2V / Stitch / Extend） |
+| **总组合时长** | 所有剪辑 ≤ 15 s |
+| **每个剪辑大小** | ≤ 50 MB |
+
+### 音频
+
+| 约束 | 值 |
+|---|---|
+| 输入方法 | URL（`http://`、`https://`）或 Base64 data URL（`data:audio/...`） |
+| 格式 | `.wav`、`.mp3` |
+| **每个剪辑时长** | `[2, 15]` s |
+| **最大剪辑数量** | 3 |
+| **总组合时长** | 所有剪辑 ≤ 15 s |
+| **每个剪辑大小** | ≤ 15 MB |
+
+参考音频仅在 R2V 变体上支持。每个条目作为 `role: "reference_audio"` 内容项转发到模型，prompt 中以 `<Audio 1>`、`<Audio 2>`、`<Audio 3>` 寻址——模型根据 prompt 框架将每个剪辑用于人声音色、音效或背景音乐。旧版单一 `audio_url` 字段映射到相同的内容形态，现在等同于传递一元素的 `reference_audio_urls`。
 
 <Warning>
-**`reference_audio_urls` cannot be the only reference input.** The model requires at least one image or video reference alongside any audio donor. Pair `reference_audio_urls` with `reference_image_urls`, `reference_video_urls`, `image_url`, or `video_url` — audio-only submissions are rejected.
+**`reference_audio_urls` 不能作为唯一的参考输入。** 模型要求在任何音频供体之外至少有一个图像或视频参考。将 `reference_audio_urls` 与 `reference_image_urls`、`reference_video_urls`、`image_url` 或 `video_url` 配对——纯音频提交会被拒绝。
 </Warning>
 
-### Request size
+### 请求大小
 
-The queue endpoint accepts JSON bodies up to **35 MB**. Inline data URLs for large videos can push past this — for multi-clip Stitch in particular, prefer URLs over inline base64.
+队列端点接受最大 **35 MB** 的 JSON 正文。大型视频的内联 data URL 可能会超过此值——尤其是多剪辑 Stitch，建议使用 URL 而非内联 base64。
 
 ---
 
-## Pricing
+## 定价
 
-Call `POST /api/v1/video/quote` to get a quote for a given request shape before submitting it to `/video/queue`. The quote endpoint is the only authoritative source; pricing details may change and shouldn't be cached or duplicated client-side.
+在提交到 `/video/queue` 之前，调用 `POST /api/v1/video/quote` 获取给定请求形态的报价。报价端点是唯一权威来源；定价细节可能会变化，不应在客户端缓存或复制。
 
-When reference video(s) are part of the request, also pass `reference_video_total_duration` (the sum of all reference clip durations in seconds) so the quote matches what `/video/queue` will charge:
+当请求包含参考视频时，还要传递 `reference_video_total_duration`（所有参考剪辑时长的总秒数），以便报价与 `/video/queue` 实际收费匹配：
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/quote \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "duration": "5s",
     "resolution": "1080p",
     "aspect_ratio": "16:9",
@@ -228,9 +221,9 @@ curl -X POST https://api.venice.ai/api/v1/video/quote \
 
 ---
 
-## Complete examples
+## 完整示例
 
-All examples assume `VENICE_API_KEY` is set in the environment.
+所有示例假设环境中设置了 `VENICE_API_KEY`。
 
 ### Text-to-video
 
@@ -239,7 +232,7 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
+    "model": "seedance-2-0-text-to-video",
     "prompt": "A golden retriever frolicking through a sunlit meadow at sunset, slow camera dolly-in, shallow depth of field, warm cinematic lighting.",
     "duration": "5s",
     "aspect_ratio": "16:9",
@@ -247,44 +240,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.0 text-to-video (4K)
+### Image-to-video（第一帧）
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-text-to-video-basic",
-    "prompt": "Ultra-detailed aerial glide over a sunlit alpine lake, crystal water, distant peaks, cinematic color grade.",
-    "duration": "5s",
-    "aspect_ratio": "16:9",
-    "resolution": "4k"
-  }'
-```
-
-### Seedance 2.5 text-to-video (longer duration)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-text-to-video-basic",
-    "prompt": "A slow aerial push over misty mountains at sunrise, clouds parting, soft orchestral ambience, cinematic widescreen framing.",
-    "duration": "20s",
-    "aspect_ratio": "16:9",
-    "resolution": "720p"
-  }'
-```
-
-### Image-to-video (first frame)
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-0-image-to-video-basic",
+    "model": "seedance-2-0-image-to-video",
     "prompt": "The lighthouse keeper turns toward the storm, lantern raised, waves crashing against the rocks.",
     "image_url": "https://example.com/lighthouse.jpg",
     "duration": "5s",
@@ -293,17 +256,17 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
 ```
 
 <Note>
-Seedance I2V models (`seedance-2-0-image-to-video-basic`, its Fast variant, and `seedance-2-5-image-to-video-basic`) **do not accept `aspect_ratio`** — the output aspect ratio is auto-derived from the input image's dimensions. Passing the field returns a 400 with *"This model does not support aspect_ratio"*. Use the T2V or R2V variants if you need explicit aspect-ratio control.
+`seedance-2-0-image-to-video`（及其 Fast 变体）**不接受 `aspect_ratio`** ——输出宽高比从输入图像的尺寸自动派生。传递该字段会返回 400 错误，并显示 *"This model does not support aspect_ratio"*。如果需要显式宽高比控制，请使用 T2V 或 R2V 变体。
 </Note>
 
-### Reference workflow — subject donor
+### Reference 工作流——主体供体
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night.",
     "reference_image_urls": ["https://example.com/character.png"],
     "duration": "5s",
@@ -312,33 +275,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Seedance 2.5 reference workflow — multi-image
-
-```bash
-curl -X POST https://api.venice.ai/api/v1/video/queue \
-  -H "Authorization: Bearer $VENICE_API_KEY" \
-  -H "Content-Type: application/json" \
-  -d '{
-    "model": "seedance-2-5-reference-to-video-basic",
-    "prompt": "Refer to <Subject 1> in <Image 1> and the style in <Image 2> to generate a 12-second clip of the same character exploring a rainy cyberpunk alley.",
-    "reference_image_urls": [
-      "https://example.com/character.png",
-      "https://example.com/style-board.png"
-    ],
-    "duration": "12s",
-    "aspect_ratio": "9:16",
-    "resolution": "720p"
-  }'
-```
-
-### Reference workflow — subject + audio donor
+### Reference 工作流——主体 + 音频供体
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Refer to <Subject 1> in <Image 1> to generate a 5-second clip of the same character walking through a neon-lit Tokyo street at night. Refer to the timbre in <Audio 1> for a soft female voiceover describing the scene.",
     "reference_image_urls": ["https://example.com/character.png"],
     "reference_audio_urls": ["https://example.com/voice-sample.mp3"],
@@ -348,14 +292,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow
+### Edit 工作流
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Strictly edit <Video 1>, changing its weather from sunny to a heavy rainstorm, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/sunny-scene.mp4"],
     "reference_video_total_duration": 5,
@@ -365,14 +309,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Edit workflow with image grounding
+### 带图像基础的 Edit 工作流
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Replace the perfume featured in <Video 1> with the face cream from <Image 1>, with all original motions and camera work preserved.",
     "reference_video_urls": ["https://example.com/perfume-ad.mp4"],
     "reference_image_urls": ["https://example.com/face-cream.png"],
@@ -383,14 +327,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Extend forward
+### 向前 Extend
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "Extend <Video 1>, generate a dramatic chase scene through narrow alleys at dusk, with neon signs flickering and rain on the pavement.",
     "reference_video_urls": ["https://example.com/alley-intro.mp4"],
     "reference_video_total_duration": 4,
@@ -400,14 +344,14 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Stitch (3 clips)
+### Stitch（3 段剪辑）
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/queue \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "prompt": "<Video 1> + a wisp of smoke transforms into a flock of birds + followed by <Video 2> + a slow dolly-in + followed by <Video 3>",
     "reference_video_urls": [
       "https://example.com/clip-1.mp4",
@@ -421,65 +365,58 @@ curl -X POST https://api.venice.ai/api/v1/video/queue \
   }'
 ```
 
-### Polling for completion
+### 轮询完成
 
-After every queue submission, save the returned `queue_id` and poll `/video/retrieve` until the response body is `video/mp4`:
+每次队列提交后，保存返回的 `queue_id` 并轮询 `/video/retrieve` 直到响应体为 `video/mp4`：
 
 ```bash
 curl -X POST https://api.venice.ai/api/v1/video/retrieve \
   -H "Authorization: Bearer $VENICE_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "seedance-2-0-reference-to-video-basic",
+    "model": "seedance-2-0-reference-to-video",
     "queue_id": "123e4567-e89b-12d3-a456-426614174000"
   }' \
   -o output.mp4
 ```
 
-The response is JSON (`{ "status": "queued" | "running" | "failed", ... }`) until the job completes, at which point the response body switches to `video/mp4` bytes. See [Video Generation](/guides/media/video-generation) for the full polling pattern.
+直到作业完成，响应都是 JSON（`{ "status": "queued" | "running" | "failed", ... }`），完成时响应体切换为 `video/mp4` 字节。完整轮询模式请参阅[视频生成](/guides/media/video-generation)。
 
 ---
 
-## Troubleshooting
+## 故障排查
 
 ### `At least one reference is required for this model`
 
-Reference-to-video submissions must include at least one of `reference_image_urls`, `reference_video_urls`, `image_references`, or `video_references`. Pure text-only generation isn't a valid R2V workflow — use a text-to-video model ID instead. `reference_audio_urls` alone is not sufficient (see the Audio section above).
-
-### Too many reference videos / images
-
-Seedance **2.0** caps R2V at **9 images** and **3 videos**. Seedance **2.5** raises those caps to **30 images** and **10 videos**. If you exceed the family limit, trim inputs or stitch offline first.
-
-### Duration / aggregate duration errors
+Reference-to-video 提交必须包含 `reference_image_urls`、`reference_video_urls`、`image_references` 或 `video_references` 中的至少一个。纯文本生成不是有效的 R2V 工作流——请改用 `seedance-2-0-text-to-video`。仅 `reference_audio_urls` 不够（请参阅上面的音频章节）。
 
-- **2.0:** per-clip reference video/audio `[2, 15]` s; combined video/audio ≤ 15 s; output 4–15 s.
-- **2.5:** per-clip reference video/audio `[2, 30]` s; combined video/audio ≤ 30 s; output 4–30 s.
+### `reference_video_urls must have at most 3 videos`
 
-Trim clips client-side before submission.
+模型将参考视频上限设为 3。如果您需要更多剪辑，先运行一次 Stitch（3 → 1），然后将输出用作后续的参考。
 
-### Prompt routes to the wrong workflow
+### `Per clip must be 2–15s` / 聚合 `> 15s`
 
-Workflow is inferred from prompt syntax. Common misroutings:
+每个剪辑时长是**含端点**的 `[2, 15]` 秒；所有参考视频的总和也上限为 15 秒。在提交之前在客户端修剪剪辑。
 
-- Wanting to **Extend** but writing `Refer to ...` → model treats your video as a *donor*, not a canvas to continue
-- Wanting to **Stitch** but writing `Refer to ...` → model picks one as the donor, ignores the others
-- Wanting to **Edit** but writing `Generate a video based on <Video 1>` → ambiguous; model may default to Reference
+### Prompt 路由到错误的工作流
 
-Use the canonical prefixes exactly as written: `Strictly edit <Video 1>, ...`, `Extend <Video 1>, ...`, `<Video 1> + ... + followed by <Video 2>`.
+工作流从 prompt 语法推断。常见的错误路由：
 
-### Face / likeness inputs rejected
+- 想要 **Extend** 但写 `Refer to ...` → 模型将您的视频视为*供体*，而非要延续的画布
+- 想要 **Stitch** 但写 `Refer to ...` → 模型挑选一个作为供体，忽略其余的
+- 想要 **Edit** 但写 `Generate a video based on <Video 1>` → 含糊；模型可能默认 Reference
 
-Public Seedance API models are **faces-off**. Do not expect a consent attestation flow. Face-bearing media may fail with a content-policy / provider error. Use the Venice app or Studio for likeness workflows.
+完全按所写使用规范前缀：`Strictly edit <Video 1>, ...`、`Extend <Video 1>, ...`、`<Video 1> + ... + followed by <Video 2>`。
 
-### Quote doesn't match the queued amount
+### 报价与队列金额不匹配
 
-If you included a reference video but didn't pass `reference_video_total_duration` to `/video/quote`, the quote and the queued amount may differ. Always pass `reference_video_total_duration` (sum of all reference clip durations, in seconds) when reference videos are present.
+如果您包含了参考视频但没有将 `reference_video_total_duration` 传递给 `/video/quote`，报价和队列金额可能不同。当存在参考视频时，始终传递 `reference_video_total_duration`（所有参考剪辑时长的总秒数）。
 
 ---
 
-## References
+## 参考
 
-- Venice video queue endpoint: [`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
-- Venice quote endpoint: [`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
-- Companion guide: [Reference to Video](/guides/media/reference-to-video) (covers Kling O3 + Grok Imagine R2V)
-- Companion guide: [Video Generation](/guides/media/video-generation) (queue / polling overview)
+- Venice 视频队列端点：[`POST /api/v1/video/queue`](/api-reference/endpoint/video/queue)
+- Venice 报价端点：[`POST /api/v1/video/quote`](/api-reference/endpoint/video/quote)
+- 配套指南：[Reference to Video](/guides/media/reference-to-video)（涵盖 Kling O3 + Grok Imagine R2V）
+- 配套指南：[视频生成](/guides/media/video-generation)（队列/轮询概览）
diff --git a/zh/guides/media/video-generation.mdx b/zh/guides/media/video-generation.mdx
index e5577571..04d1731e 100644
--- a/zh/guides/media/video-generation.mdx
+++ b/zh/guides/media/video-generation.mdx
@@ -340,7 +340,7 @@ await fetch(`${BASE_URL}/video/complete`, {
 报价以美元为单位。
 
 <Info>
-**报价参数仅用作定价输入。** 您发送到 `/video/quote` 的字段——包括 `aspect_ratio`、`audio` 和 `reference_video_total_duration`——仅用于计算价格。它们**不会**被转发到 `/video/queue`，也不会影响生成的视频。即使某些模型（例如 `seedance-2-0-image-to-video-basic`）在生成时会拒绝 `aspect_ratio`，向 quote 发送 `aspect_ratio` 也是可接受的，因为 image-to-video 模型会从输入图像推导出输出宽高比。
+**报价参数仅用作定价输入。** 您发送到 `/video/quote` 的字段——包括 `aspect_ratio`、`audio` 和 `reference_video_total_duration`——仅用于计算价格。它们**不会**被转发到 `/video/queue`，也不会影响生成的视频。即使某些模型（例如 `seedance-2-0-image-to-video`）在生成时会拒绝 `aspect_ratio`，向 quote 发送 `aspect_ratio` 也是可接受的，因为 image-to-video 模型会从输入图像推导出输出宽高比。
 </Info>
 
 ### 参考视频到视频的定价