How does transcription work in Canto?

When you stop a recording, Canto uploads the audio, transcribes it in the cloud, and then generates a title and notes. The transcript is not live — it appears a few minutes after the recording ends, split into speaker turns with a timestamp on every word.

What happens after I stop recording?

1. Upload — the audio is compressed and sent to your Canto library. On mobile this happens in the background; keep the device online until the upload finishes.
2. Transcription — the audio is transcribed with the spoken language detected automatically and speakers separated into turns. See Which languages does Canto transcribe? and How does Canto tell speakers apart?
3. AI preparation — Canto writes a title and notes, and indexes the recording so chat and search can find it. See How do AI notes work?

Once all three finish, the recording shows Ready.

What do the statuses mean?

• Recording — audio is still being captured on your device.
• Processing — the recording has stopped and the audio is being prepared for upload.
• Uploading — the audio is on its way to your library.
• Queued — uploaded and waiting for transcription to start.
• Transcribing — the transcript is being generated.
• Preparing AI — the transcript exists; the title, notes, and search index are being built. You can already read the transcript at this point.
• Ready — everything is done.
• Failed — transcription didn't complete. See Retry a failed transcription.

A recording can also show Locked or Not transcribed on the free plan — that's a quota state, not an error. See Why is my recording locked or not transcribed?

Is transcription live while I record?

No. Canto records first and transcribes after the recording stops and uploads. During a long recording you see the timer and can add bookmarks, but there is no on-screen text until transcription finishes.

How long does it take?

Transcription runs much faster than real time — a one-hour recording is typically ready in a few minutes after the upload completes. Upload time depends on your connection; the audio is compressed to roughly 14 MB per hour. If a recording sits in Transcribing far longer than that, see My transcription is stuck or failed.

What does the transcript contain?

The transcript is a sequence of speaker turns, each labelled with a speaker name and a start time. Every word carries its own timestamp, so you can tap anywhere to play the audio from that moment — see Jump the audio to any word. Transcripts sync to all your devices, and you can correct mistakes afterwards — see Edit a transcript.