MetaMuse Voice
Muse Voice Transcribe 1.0
Muse Voice Transcribe handles live streams and recorded audio. It supports 25 languages, speaker labels and vocabulary hints.
What the release provides
The API returns turn-level timestamps rather than word-level timestamps. Modes cover push-to-talk, model-detected turn boundaries and speaker diarization. It does not synthesize speech.Muse Voice Transcribe guide
How to use it well
Practical guidance. Choose a turn mode that matches how people use the microphone. Do not assign a speaker's real identity from an anonymous speaker label alone. Keep the recording available for review when names or technical terms affect the result.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | muse-voice-transcribe-1.0 |