Record or upload audio and see how Omi's medical speech-to-text
transcribes it — right in your browser, with optional custom vocabulary.
♪
Drop an audio file here, or click to browse
wav, mp3, m4a, ogg or flac · up to 25 MB
0:00 / 10:00
Recording…
Tap to start recording. Max 10 minutes.
Ready for a live session
A live draft appears as you speak. When you stop, it's replaced by the final transcript.
0:00·Listening
ⓘPreview sessions are capped at 30 minutes. The final transcript follows shortly after you stop.
1 · Live draftAppears as you speak
2 · CorrectedRefinements replace the draft in real time
3 · Final transcriptReady after you stop
Live draft
Working text only · never use for documentation
Start a session and speak to see the live draft.
✓ Draft and corrected text replaced by the final transcript.
Microphone stopped. Building the final transcript while this connection stays open…
The playground runs in English. Other languages are available through the
API.
Labels decorate corrected and official text; they never change the words.
Your clinical terms
Add terms likely to be spoken in this encounter.
Try:
0 / 20 termsTerms apply to corrected and official text, not the live draft.
Audio is used only to create the transcript — it isn't stored.
Transcript
Output is an AI-generated draft and may contain errors — it is not a
medical record and must not be used for clinical decisions. By using this
preview you agree to the
Preview Terms. Learn how the
production API keeps patient data in region in
Speech-to-Text models.