⚡️ Markdown Translator

English | 中文

Translate Markdown & preserve every format — headings, code, formulas, all intact

MD Translator solves the broken-formatting problem that plagues Markdown translation. It delivers high-quality translations while accurately preserving every Markdown construct — code blocks, LaTeX formulas, FrontMatter metadata, links, and emphasis all stay intact. Connect to 7 traditional translation APIs (DeepL, Google, Azure, DeepLX, Qwen-MT, TranslateGemma, GTX) or 17+ LLM providers — 25+ engines in total — and translate into 120+ languages simultaneously. Everything runs locally in your browser: your source never leaves your machine and API keys stay in local storage.

👉 Try it online: https://tools.newzone.top/en/md-translator

Key Features

• Format-Preserving: Tokenize FrontMatter, code blocks, LaTeX, links, image paths, headings, lists, blockquotes, and HTML/JSX tags into placeholders; restore them losslessly after translation.
• Native Markdown Support: Translates only the prose layer; headings, lists, code blocks, links, emphasis, and LaTeX stay byte-perfect. Full CommonMark + GFM (tables, task lists, strikethrough) support, plus opaque-block handling for MDX and Astro components.
• Plain Text Mode: Toggle "Ignore Formatting" to skip format parsing for plain text inputs (TXT, HTML, logs) — useful when the Markdown tokenizer would over-protect, or for complex MDX you want to translate verbatim.
• Batch File Upload: Drop a whole docs/ directory and translate every file in one click — each translated file exports separately, ready to drop back into Hugo, Jekyll, Hexo, VitePress, or Docusaurus i18n folders.
• Multi-Language Output: Translate to 120+ languages in one pass — each language exported as its own file.
• Context-Aware Translation (LLM only): Surrounding paragraphs included as context for better coherence and terminology consistency. Custom system / user prompts let you lock in terminology and style.
• RTL Language Support: Automatically detects and adjusts text direction for RTL languages such as Arabic, Hebrew, Urdu, and Persian.
• Unlimited Caching (IndexedDB): All translations cached locally with no browser-storage size limit.
• Text Extraction: Strip Markdown syntax to clean plain text for summarization, NLP, or search indexing.
• Runs Locally / Privacy-First: All reading, parsing, and translation happen in your browser. LLM requests go straight from your browser to your configured endpoint; source content never touches our servers and API keys live only in local storage.
• Multi-Locale UI: Powered by next-intl, with full UI translation across 18 languages.

Supported Markdown Elements

Element	Syntax	Protected
FrontMatter metadata	--- block	Optional
Headings	# … ######	✅
Lists / task lists	- / * / 1. / - [ ]	✅
Tables	| col | col |	✅
Blockquotes	> quote	✅
Links & image paths	[text](url), 	✅
Emphasis	**bold**, _italic_, ~~del~~	Inline
Code blocks / inline code	``` and `	Optional
Inline / block LaTeX	$formula$, $$formula$$	Optional
HTML / JSX & MDX components	<span>, <br/>, <Alert>	✅

FrontMatter, code blocks, and LaTeX formulas can be translated or kept as-is — each is an independent toggle. MDX and Astro component tags are protected as opaque blocks while plain text between them is translated normally.

Translation APIs

Supports 7 traditional MT APIs and 17+ LLM providers:

Traditional APIs

API	Quality	Stability	Free Tier
DeepL	★★★★★	★★★★☆	500K chars/month
Google Translate	★★★★☆	★★★★★	500K chars/month
Azure Translate	★★★★☆	★★★★★	2M chars/month (first 12 months)
DeepLX (Free)	★★★★☆	★★★☆☆	Self-host or free public endpoints
Qwen-MT	★★★★☆	★★★★☆	Alibaba DashScope quota
TranslateGemma	★★★★☆	★★★★☆	Self-host (LM Studio / Ollama / etc.)
GTX API (Free)	★★★☆☆	★★★☆☆	Free (rate-limited)

LLM Providers

Supports DeepSeek, OpenAI, Claude, Gemini, Qwen, Moonshot, Doubao, Zhipu GLM, MiniMax, Mistral, Perplexity, Cohere, OpenRouter, Groq, SiliconFlow, Nvidia NIM, Azure OpenAI, plus any Custom (OpenAI-compatible) endpoint (Ollama / LM Studio / vLLM / Together AI / Fireworks AI etc.). Each provider has a configurable model list, temperature, system / user prompts, and per-request thinking-mode toggle.

Context-Aware Translation (LLM only)

LLM modes can send surrounding lines as context for each batch, improving paragraph-level coherence and terminology consistency.

• Concurrent Lines: max lines translated in parallel (default 20). Too high triggers rate limits.
• Context Lines: lines included per batch as context (default 50). Higher = better coherence but more tokens.

⚠️ Caveat: Markdown is complex — enabling context mode may slightly raise the risk of formatting errors (unclosed code blocks, list indentation drift). Spot-check output, especially for documents with deeply nested structure.

Use Cases

• 📚 Translate multilingual technical documentation in bulk
• 🌐 Open-source project documentation i18n (VitePress / Docusaurus i18n.locales)
• 📄 Localize a GitHub README or an entire doc site, then drop files straight back into the source tree
• ✍️ Bilingual Markdown blog content sync (Hugo / Jekyll / Hexo)
• 🧮 Format-preserving translation of mixed content (text + code + formulas)
• 🔍 Strip Markdown to plain text for summarization / NLP / search indexing

FAQ

Which engine should I use for technical docs? AI/LLM is strongly recommended — models recognize library names, function names, and variables in context without mistranslating them. Claude Sonnet leads on API-doc terminology accuracy, DeepSeek delivers excellent value, and Gemini fits book-length documentation thanks to its long context. Traditional machine translation is best reserved for quick previews.

How are code blocks and formulas kept intact? A placeholder-protection strategy: code fences, inline code, LaTeX ($...$, $$...$$), link URLs, image paths, and HTML/JSX tags are swapped for placeholders (e.g. <<<MULTILINE_CODE_x>>>) before translation, then restored verbatim. FrontMatter is skipped by default; separate toggles control FrontMatter, code blocks, LaTeX, and link text.

Does it support GFM, MDX, or Astro? Standard CommonMark and GFM (tables, task lists, strikethrough, fenced code) are fully supported. For MDX and Astro, component tags are treated as opaque blocks while plain-text content between them is translated; flip "Ignore Formatting" to translate complex MDX as plain text.

Is my content uploaded to a server? No. Reading, parsing, and translation all run client-side. LLM requests go directly from your browser to your configured API endpoint, and API keys are stored only in local browser storage. Translation cache uses IndexedDB.

What about subtitles or JSON config files? Use the companion tools — Subtitle Translator (SRT/ASS/VTT/LRC) and JSON Translator (i18next / next-intl / vue-i18n) — which share the same engine configs and API keys, so no re-setup is needed.

Tech Stack

• Framework: Next.js 16 (App Router) + React 19 with the React Compiler
• UI: Ant Design 6 + Tailwind CSS 4
• i18n: next-intl
• Caching: idb (IndexedDB)
• Testing: Vitest — restorePlaceholders and other placeholder utilities ship with unit tests

Getting Started

Requirements

• Node.js >= 20.9.0
• Yarn (recommended), npm, or pnpm

Install & Run

git clone https://github.com/rockbenben/md-translator.git
cd md-translator

yarn install
yarn dev

Visit http://localhost:3000.

Production Build

yarn build

Documentation & Deployment

For detailed configuration, API setup, and self-hosting instructions, see the Official Documentation.

Quick Deployment: Deploy Guide

Contributing

Contributions are welcome! Feel free to open issues and pull requests.

1. Fork the repo and create a feature branch
2. Run yarn and yarn dev locally
3. Add tests / docs when applicable
4. Submit a PR with a clear description

License

MIT © 2025 rockbenben. See LICENSE.