Features
Speaker diarization
Automatically work out who spoke when, and label the transcript so a meeting reads as a dialogue.
Diarization is the step that turns a wall of text into a conversation. The engine separates voices in the recording and labels each stretch of speech, so you can see who said what without listening back.
Labels start generic — Speaker 1, Speaker 2 — and you rename them once. The transcript, the summary, and every export update together.
00:14MayaRevenue grew fifteen percent this quarter.
00:31DanielWhat are we committing to for Q2?
Up to six speakers
Covers most meetings, interviews, panels, and depositions.
Renameable labels
Rename once and the change propagates through the transcript and every export.
Included on every plan
Free tier included. Several competitors charge for this separately.
Better with separate tracks
If your recorder captures per-participant tracks, accuracy improves substantially.
Questions
What people ask about speaker diarization before they rely on it.
What happens with more than six speakers?
The transcript stays complete and accurate; the labels become less reliable. Large panels read better as plain text than as attributed dialogue.
Can I rename Speaker 1 to a real name?
Yes, once. The change flows through the transcript, the summary, and every export.
Why do labels blur when people interrupt each other?
Overlapping speech is genuinely ambiguous — two voice signals occupy the same moment. Every diarization system degrades here, ours included.
Does diarization recognise the same person across different recordings?
No. Speaker labels are worked out fresh for each recording and don’t carry over — the same person in two separate meetings gets labelled independently in each. We don’t build or keep a voice profile that identifies someone across sessions.
Be first in when it opens.
Sign-ups are not open yet. Leave your address and you get one email the day they are — sixty minutes free, no card.