Uploaded recordings are now separated by voice, so a panel or interview reads as a conversation instead of one undifferentiated block. Speakers arrive as Speaker 1, Speaker 2 and so on - name them once and the names apply across the transcript, the subtitle exports, and anyone you share the link with. Separation is automatic and imperfect, so the whole transcript is editable by hand.
- Uploads are diarized: each line is attributed to a speaker, colour-coded and grouped so a run of consecutive lines sits under one namenew
- Name each speaker inline - names are saved with the transcript and used everywhere it appearsnew
- Add a speaker the automatic pass missed, or remove one (its lines stay, unattributed)new
- Edit the transcript by hand: change any line's text, timestamp or speaker, split a line in two at the cursor, insert lines, delete linesnew
- Download subtitles as SRT or WebVTT, with or without speaker names baked into each captionnew
- Speaker separation applies to uploaded files - live stream transcripts are unchangedimprovement