# transcribed.dev

> Transcribe any YouTube video — even without captions. Search your library, send transcripts to an LLM, or pull them via API.

## API

Base URL: https://transcribed.dev
Auth: Bearer token via `Authorization` header. Keys start with `ytt_sk_`.

### Transcribe a Video (async)

POST /api/transcripts
- Body: `{"url": "YOUTUBE_URL"}`
- Returns 202 with `{"id", "videoId", "status": "processing"}`
- Poll GET /api/transcripts/:id every 2-3 seconds until `status` is `"done"`

### Get Transcript

GET /api/transcripts/:id
- Returns JSON with title, author, videoUrl, transcript (JSON segments), source, status
- Append `?format=text` for plain text (recommended for LLM input)

### List Transcripts

GET /api/transcripts
- Optional: `?q=SEARCH_TERM`

### Delete Transcript

DELETE /api/transcripts/:id

## Error Codes

- 202: Processing — poll until done
- 400: Invalid or unsupported URL
- 401: Missing or invalid API key
- 402: Quota exceeded — response includes `upgradeUrl`
- 404: Video not found or unavailable
- 429: Rate limited — wait 30-60s

## Notes

- Submitting the same URL twice returns the cached transcript (no re-processing, no quota charge)
- `source` is `"youtube_captions"` or `"whisper"`
- Supported formats: youtube.com/watch?v=, youtu.be/, youtube.com/shorts/, youtube.com/embed/
- Free tier: 25 transcripts/month. Pro ($9/mo): 1,000/month. Power ($19/mo): 5,000/month.

## Docs

- [API docs](https://transcribed.dev/docs/api)
- [Getting started](https://transcribed.dev/docs/getting-started)
- [MCP server](https://transcribed.dev/docs/mcp)
- [Developers](https://transcribed.dev/developers)
- [Pricing](https://transcribed.dev/pricing)