Meetings · Voice Synthesis · guide

Click ● Start recording on the Meetings page (web, iPhone, iPad, or the Mac/Windows desktop app). Pick the meeting type, attendees, project. Hit start. While you're recording you get a live, browser-native transcript with best-effort speaker tags, and can turn on Live extract to send transcript chunks out every 12 seconds for in-progress action items, decisions, and risks. The accurate, fully diarized transcript and extraction pass runs after you hit Stop & process (occasionally in the background for longer recordings) — that's the version that becomes the finalized minutes.

For an online call (Teams, Zoom, Google Meet, etc.): use the Mac or Windows desktop app. It captures the call's system audio directly, so it hears everyone on the call regardless of headphones or speaker volume. The web and iPhone/iPad recorders only capture your microphone — great for in-person meetings, but they won't reliably pick up an online call's audio.

The first time the diarizer hears a speaker, you'll see speaker_3 in the transcript. Tap and assign once; we remember their voice for future meetings on this project.

Click Finalize. Action items become Tasks (assignees notified). Decisions and risks get tagged with the RFI, CO, submittal, PO, drawing, or spec they reference, so you can follow up in that tracker yourself. Owner-side users see the redacted view if visibility is "team_plus_owner."

Every extracted item has an AnchorChip showing the audio timestamp. Click → audio scrubber seeks to that moment, plays for the spanned range, then pauses.