Four models. One API.
Speech-to-text for recordings and live audio, speaker diarization and speaker ID, built to stay accurate in real conversations.
Async speech-to-text
a5s-async-v1
A5S Async
Transcripts with speaker labels for meetings, calls, interviews and media.
29% fewererrors than the next best on 5 meeting benchmarks
$0.18/h
Streaming speech-to-texta5s-small-v2-streaming
A5S Streaming
Real-time speech-to-text for live captions and voice agents.
26.27% WERon noisy, far-field audio, ahead of 8 cloud systems
$0.30/h
Speaker diarizationaircaps-diarization-v1
Diarization
Who spoke when, across meetings, calls and broadcasts.
12.03% DERlowest of 11 systems across 7 datasets
$0.08/h
Speaker identificationaircaps-speaker-id-v1
Speaker ID
Recognize enrolled voices through noise and music.
3.11% EERwith noise louder than speech
$0.02/h
Which one?
A5S Async and A5S Streaming are on every account. Diarization and Speaker ID are enabled on request.
| You have | Use |
|---|---|
| Recordings, uploads, meeting files | A5S Async |
| The most accurate speaker-labeled transcript of a recording | A5S Async |
| A live microphone or call that needs captions now | A5S Streaming |
| Live audio that needs speaker labels as it happens | A5S Streaming with live speaker labels |
| Only who spoke when, or speakers for a transcript you already have | Diarization |
| Which known person is speaking: names instead of speaker numbers | Speaker ID |
Hear what other models miss.
Run AirCaps on your own audio. Every account includes 5 free hours.