Client calls
Record the call, get a speaker-labeled transcript, ask the local model for a follow-up email draft based on the commitments.
Record the call, get a speaker-labeled transcript, ask the local model for a follow-up email draft based on the commitments.
Five voices, one summary. Action items extracted by speaker, ready to paste into Linear or Jira.
Subject + interviewer labeled separately. Quote-search the transcript later without re-listening.
No. The recording is captured locally and transcribed locally with Whisper. The audio file lives in a local SQLite database.
Diarization groups voices into clusters automatically. Typilot labels them speaker-1, speaker-2 by default; you can rename each speaker once. From then on, Typilot recognises that voice in future meetings and applies the name automatically - accuracy compounds the more you record with the same team.
Yes. Configure Typilot to keep audio or delete it after the summary is generated.