Meeting recording captures your microphone and your computer's audio output, so both sides of a call are transcribed. It runs from the tray menu — Start Meeting Recording — rather than a hotkey.
Capturing the other side
| Platform | Method | Extra setup |
|---|---|---|
| Windows | Loopback capture | None |
| macOS | ScreenCaptureKit | Screen Recording permission required |
| Linux | PulseAudio monitor | None |
Recording legality varies by jurisdiction and often requires the consent of everyone present. That is your responsibility, not the app's.
Telling speakers apart
Speaker identification is on by default and runs entirely on your machine — it needs a one-time model download. The transcript is then split by speaker rather than presented as one undifferentiated block.
Transcribing existing files
Use Transcribe a File… from the tray, or the button on the History page, to run an existing recording through the same pipeline. This depends on ffmpeg — bundled on Windows and macOS, and taken from your system PATH on Linux.
Storage
Meeting recordings have their own Keep recorded audio setting and retention window, including "keep forever" — separate from dictation history. Transcripts can be exported as VTT, SRT, plain text, Markdown or JSON.