Turn a meeting recording into minutes
Choose a meeting recording and it is transcribed with speakers distinguished. The audio never leaves your device, so recordings that cannot go outside your organisation can be processed as they are. Speaker separation is enabled from the start.
If you know the speaker count, setting it greatly improves accuracy
Getting usable minutes
This page enables speaker separation from the start, because a meeting involves several voices and minutes are hard to use unless you can tell who said what. If you know how many people took part, setting the speaker count improves separation further. For recordings heavy with proper nouns or jargon, switching to the small model or larger reduces dropped words.
How the meeting was recorded matters too. Two things help most: every voice captured at a similar level, and people not talking over each other. Recordings made with an online meeting tool's built-in recorder tend to separate better than a single microphone in the middle of a conference table.
Frequently asked questions
How many participants can it handle?
Up to eight speakers can be distinguished. If you know the number, set it in the options. Otherwise it is estimated automatically, though participants who speak only briefly may be missed.
Could the meeting audio be sent outside?
No. Both transcription and speaker separation run inside your browser, and neither the audio nor the resulting text is sent to any external server. Confidential meetings are safe to process.
Can I use real participant names for the speakers?
Yes. You can name each speaker on screen, and merge speakers if one person was split into two. Individual lines can be reassigned as well.
Does it work with online meeting recordings?
Yes. Pick the recording file from Zoom, Teams, Google Meet and so on (MP4 or M4A) directly. For video files, only the audio track is extracted.
Getting the audio ready, and after transcribing
Help us improve the service
Sending us the circumstances of a run — successful or failed — helps us track down problems (a failure rate needs both). Knowing which devices and settings fail most often lets us revise the recommended defaults and fix the bugs behind them. If you don't mind, leaving this on is a real help.
What is sent
Your device specs (GPU model, memory, core count), browser and OS, the model and language you chose, how far the run got, the error details, the file's format with a rough size and length, and how long it took.
What is never sent
Audio, transcripts, file names, and IP addresses. The data goes only to this site's own server, never to a third-party service.