v0.6.0-beta7
Voice samples in the “Other voices” panel got a real polish pass. You can now actually recognize who you’re hearing — instead of guessing and reaching for the play button again.
Voice samples that let you recognize the voice
Before: you clicked play next to Speaker 2, heard a half-second “yeah,” and were no closer to knowing who that was. The button was playing the speaker’s first segment, which for any non-primary speaker is often a back-channel utterance — too short to matter no matter what we capped the length at.
Now: the button plays the speaker’s longest segment instead, capped at 15 seconds. If they spoke for a real stretch in the meeting, you hear it.
You can see the row is alive
A 2pt progress line now appears under the speaker row while a sample plays, filling left-to-right in that speaker’s own color. Solves the awkward moment when a Whisper segment spans a natural pause and the audio goes briefly quiet — you can see the row is still working through the clip and the voice is coming back.
The stop button is also filled in the speaker’s color now, instead of a quiet outline. The “currently playing” state reads at a glance.
Only one row plays at a time
Click Speaker 2, click Speaker 3 mid-playback, and Speaker 2 stops cleanly the instant Speaker 3 starts. No more accidental overlapping audio in the panel. You can sample your way through every speaker in a meeting at whatever pace you want.
Carried over from beta6
- AirPods plugged in after launch record cleanly (engine input-binding fix)
- Library sidebar can’t be pushed off-screen on narrow windows (hard window minWidth)
- “Other voices” panel is toolbar-toggleable + doesn’t reflow mid-render
- installTap silent-hang fix from beta5 still in place
Nothing changed in
- WhisperKit transcription pipeline
- VBxHMM clustering algorithm
- People model + file formats
- License code format
Auto-update
Existing beta channel testers get this within 24 hours. Or trigger now: Minutiva Private → Check for Updates…