Transcripts and captions
Live captions on your Mac, the cloud transcript that replaces them, spoken language, and custom words.
Every recording can have two kinds of text: live captions made on your Mac while you record, and a cloud transcript made after it uploads.
Live captions
The Mac app writes captions on your Mac as you speak. They are on by default and work without an account. They need the Speech Recognition permission, and you can turn them off under Settings → Live captions in the Mac app.
Talk uses the same on-device recognition for the text it pastes.
Cloud transcript
When you're signed in, recordings are also transcribed in the cloud. The cloud transcript replaces the live captions once it's ready. It's the complete record the brief is written from.
- On the Mac, audio uploads in segments while you record, so text can arrive in batches before you stop.
- In the browser, audio uploads in segments of about four minutes.
Meetings are recorded as two tracks, so lines are labelled You for your microphone and Them for the call's audio. Eavesy does not tell individual people apart within a track.
Spoken language
Transcription is set to English by default. To change it, open the web app and go to Settings → Transcription → Spoken language. Pick your language, or choose to detect it automatically. A fixed language is more accurate when you always speak the same one.
Custom words
Add names, products and jargon under Settings → Transcription → Custom words, one per line or separated by commas. You can add up to 100. They help both cloud transcripts and on-device captions on a signed-in Mac spell those words your way.
Next
These docs are open source. Spotted something wrong or missing?