Skip to content

🗣️ docs: Document gpt-transcribe Hints and Live Dictation - #831

Open
Nk-rodrigues wants to merge 1 commit into
LibreChat-AI:mainfrom
Nk-rodrigues:docs/gpt-transcribe-live-stt
Open

Nk-rodrigues wants to merge 1 commit into
LibreChat-AI:mainfrom
Nk-rodrigues:docs/gpt-transcribe-live-stt

Conversation

@Nk-rodrigues

Copy link
Copy Markdown

Summary

Documents the speech-to-text additions in LibreChat-AI/LibreChat#17190, which closes the gap described in LibreChat-AI/LibreChat#14503. The speech config reference now covers speech.stt.openai.prompt, keywords and languages, and which models receive each. It also covers the new speech.stt.openai.realtime block (model, url, delay) and the realtime value of speechTab.speechToText.engineSTT, with its fallback to external or browser.

The Speech Settings page gains a gpt-transcribe hints example and a Live dictation section. That section covers how users pick Live, its requirements (a single openai STT provider), the server-side client secret, allowedAddresses and proxy handling, the fact that live transcripts bypass filters transcript rules, per-minute billing, and an Azure OpenAI example that uses the openai provider.

LibreChat-AI/LibreChat#17190 is still open, so every new field is marked Pending in the existing style, with a hidden comment on each page to remove the markers once it merges.

Change Type

  • Documentation update

Testing

  • All YAML examples on both pages that touch speech.stt or speechTab parse with configSchema from the LibreChat PR branch. The two exceptions are the existing local Whisper examples, which omit apiKey and were already failing before this change.
  • The Azure OpenAI example matches the configuration I tested against a live Azure OpenAI resource for 🗣️ feat: First-Class gpt-transcribe and Live gpt-live-transcribe Speech-to-Text LibreChat#17190. Recorded dictation used the deployment URL, and live dictation used the /openai/v1 realtime URL.
  • pnpm lint:prettier, pnpm typecheck and pnpm lint pass.
  • pnpm build passes.

Checklist

  • Defaults, ranges and behavior checked against the LibreChat source, not memory
  • Only English pages edited; translations come from CI
  • No new pages or nav entries; the existing config reference and Speech Settings page were updated

Documents the speech-to-text additions in LibreChat-AI/LibreChat#17190 (issue #14503), marked Pending until it merges:

- `speech.stt.openai.prompt`, `keywords` and `languages`, and which models receive each.
- `speech.stt.openai.realtime` (`model`, `url`, `delay`) for live dictation over the Realtime API, with its requirements, security model, content-filter and billing notes, and an Azure OpenAI example.
- The `realtime` value of `speechTab.speechToText.engineSTT` and its fallback.
@vercel

vercel Bot commented Oct 16, 2026

Copy link
Copy Markdown

@Nk-rodrigues is attempting to deploy a commit to the LibreChat's projects Team on Vercel.

A member of the Team first needs to authorize it.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant