What it does
Whisper API transcribes audio and video with the open-source Whisper speech model, which the service runs on its own infrastructure, so you do not need an OpenAI key. You can send a file or a remote URL through the API, or use the browser dashboard with drag-and-drop and live progress. Transcripts export as JSON, text, VTT, DOCX or PDF, and the API example also shows SRT output. You pick a Whisper model size to trade speed against accuracy, and other model parameters can be tuned. The site lists 98+ languages with automatic language detection and speaker diarization, and says files can be up to 10GB with a Pro subscription. Uploaded files are deleted after 24 hours and only the transcript is kept, and the site says your content is never used to train models. Billing is by API credits. A job's cost depends on model size, audio length (billed per started hour), diarization and file size. Credits do not expire and there are no monthly fees.
Capabilities
- Transcribes audio and video from file uploads or remote URLs
- Choice of Whisper model sizes, from faster to more accurate
- Exports to JSON, text, VTT, DOCX and PDF
- Speaker diarization and automatic language detection
- Browser dashboard for people who do not write code
- Uploaded files are deleted after 24 hours
When to use it
- You want transcripts through an API without running Whisper yourself
- You prefer pay-as-you-go credits to a monthly plan
- You need subtitle-style output such as VTT from recorded audio or video
