What Amber detects automatically
Tempo, key, instrument, transcript, and title — every recording is tagged after the fact, with no setup required.
The moment you stop recording, Amber starts working out what's actually in the audio — completely on its own, with nothing for you to trigger or configure. A recording is usable and playable right away; the tags fill in a short while after, in the background.
What gets detected
- Tempo (BPM) and key — worked out from the audio itself using on-device signal analysis, the same kind of technique a DJ or DAW's auto-tempo feature uses.
- Instrument — guitar, piano, voice, drums, and other common categories, recognized from the sound itself.
- Transcript — if you talked during the recording (a scratch vocal, a note-to-self about the take), Amber transcribes it in full.
- Title — a real name for the recording instead of a timestamp, generated from what was detected (and from the transcript, when there is one).
Why it isn't instant
Detection runs as a background job after the recording is saved, not while you're recording. That's deliberate — it means analysis can never slow down or interrupt capture, even on an older device or when several recordings are queued up at once. Give it a little time; reopening the recording later (or pulling down to refresh your library) will show the finished tags.
When a tag looks wrong
Automatic detection is a best guess, not a certainty — see Fixing a wrong tag in this section for how to correct one, and why your correction sticks for good.
Was this helpful?
Your feedback shapes what we write next.