# typeheard

> Upload audio or video, get a transcript (txt, md, srt, vtt, json). whisper.cpp on our own hardware.

- POST https://typeheard.com/api/v1/transcripts  multipart: file, language (auto|en|es|...), tier (auto|preview|full), title
  -> 202 { id, status, status_url, url }
- GET  https://typeheard.com/api/v1/transcripts/{id}         poll until status is "done"
- GET  https://typeheard.com/api/v1/transcripts/{id}/{txt|md|srt|vtt|json}
- Free: the first 3 minutes of any file, no key.
- Whole file: Authorization: Bearer <api key> (1 minute of credit per audio minute), or pay per call with x402.
- MCP: npx -y @profullstack/typeheard-mcp   CLI: npx -y @profullstack/typeheard <file>
- Docs: https://typeheard.com/docs