Skip to main content

Voice, speech, and audio

Everything audio lives in Settings → Audio: speech-to-text, text-to-speech, voice mode, and voice memos.

Speech-to-text​

Any OpenAI-compatible transcription API. Defaults: base URL https://api.openai.com/v1, model whisper-1. One STT config powers three things:

  • transcribing spoken input in chat and voice mode
  • transcripts for voice memos
  • voice notes sent to messaging bots

Long recordings are automatically chunked before transcription (this path needs ffmpeg on the machine).

Text-to-speech​

Off by default. Any OpenAI-compatible /audio/speech endpoint; defaults model tts-1, voice alloy, mp3 output. Responses are read aloud sentence by sentence as they stream, so playback starts before the response finishes. Playback speed is adjustable from 0.5× to 2×.

Voice mode​

Hands-free conversation: you talk, the AI answers, the reply is read back, and the mic re-arms for your next turn. The STT source is configurable:

  • browser: the browser's built-in speech recognition, no API key needed
  • provider: the STT API configured above

Voice mode has its own customizable system prompt, separate from your chat prompts. This is useful for keeping spoken answers short.

Voice memos​

Record audio from any device and it's saved into the workspace as a file. When speech-to-text is configured, the memo is transcribed to markdown automatically: dictate a review note from your phone and read it in the editor at your desk.

This content is for informational purposes only and does not constitute a warranty, guarantee, or contractual commitment. Open WebUI is provided "as is." See your license for applicable terms.