Reverts speech-based fallback trigger. Meeting mode now strictly
requires system audio energy > 0.005 to trigger auto-start.
The system_queue always has data (even silence) from ScreenCaptureKit,
so checking queue emptiness caused false positives. Speech alone
must never trigger meeting mode.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Meeting mode was triggering on any speech (VAD) which caused false
starts when user spoke near the mic without being in a call. Now
requires system audio energy above threshold as primary trigger.
Speech is only used as fallback when combined with ANY system audio
data present, catching low-volume Electron/WebRTC apps.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Meeting mode auto-start now triggers on system audio energy OR speech
detection (VAD). Fixes issue where Electron-based apps like Nextcloud
Talk were not triggering recording because ScreenCaptureKit didn't
capture their audio reliably. Also improved debug logging.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
3CX on macOS is a Chrome PWA (com.google.Chrome.app.dmmmfdbaapefjmjeomigbhpmjbbjgnph),
not a native app. The previous allowlist only checked for "3cx" in the bundle ID,
which never matched. Now recognizes the known PWA ID and protects the Chrome main
process from exclusion when 3CX is running, since ScreenCaptureKit routes PWA audio
through Chrome.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add MP3 conversion and chunking for files >18MB
- Support meetings up to 2+ hours
- Add email button directly in history view
- Implement custom logo upload in Settings for white-labeling
- Add read_image_as_base64 Rust command
- Update README with new features and ffmpeg requirement
