Turn a recording into text with NVIDIA's open speech model: 40 languages detected automatically, optional word timestamps and speaker labels, or translation to or from English. Bring your own apimodels.app API key; it costs $0.0025 per minute of audio and failed jobs are not charged.
No key yet? Sign up at apimodels.app and create one. This page calls api.apimodels.app straight from your browser; your key never passes through any other server.
Uploads from this page: audio up to 10 MB, video up to 100 MB. A public link can be up to 60 minutes and 200 MB.
| Audio length | Price |
|---|---|
| 1 minute | $0.0025 |
| 10 minutes | $0.025 |
| 60 minutes (maximum per request) | $0.15 |
Billed by the second of audio, measured when the task is created. Timestamps, speaker labels and translation cost nothing extra. Result files are kept for 30 days.