# Media Caption

> Credit-billed REST API for retrieving public YouTube transcripts, with single and bulk transcript jobs, AI transcription of uploaded or URL-referenced media, job polling, and signed job-level webhooks. Base URL https://api.mediacaption.io/v1; authenticate with `Authorization: Bearer mc_live_xxx` or `X-API-Key`.

Generated by API Evangelist on 2026-09-03 from the provider's public docs and OpenAPI contract; the provider serves no /llms.txt of its own.

## Docs

- [API overview](https://www.mediacaption.io/docs): Media Caption API v1 documentation home
- [Quickstart](https://www.mediacaption.io/docs/quickstart): create a key, first transcript request, production checklist
- [Authentication](https://www.mediacaption.io/docs/authentication): bearer mc_live_ keys or X-API-Key; bearer takes precedence
- [OpenAPI 3.1 contract](https://www.mediacaption.io/docs/openapi/mediacaption.v1.yaml): 13 operations plus 3 declared webhooks
- [Webhooks](https://www.mediacaption.io/docs/webhooks): job.completed, job.completed_with_errors, job.failed; HMAC-SHA256 signatures
- [Billing](https://www.mediacaption.io/docs/billing): credits are frozen at acceptance, charged on success, released on failure
- [Data retention](https://www.mediacaption.io/docs/data-retention): transcript bodies expire after 3 days
- [Errors](https://www.mediacaption.io/docs/errors): stable machine-readable codes in an { error: { code, message } } envelope
- [Rate limits](https://www.mediacaption.io/docs/rate-limits): 60 rpm general, 10 rpm single transcript, 5 concurrent bulk jobs
- [Changelog](https://www.mediacaption.io/docs/changelog): dated documentation and contract changes

## API operations

- GET /user — account profile
- GET /balance — credit balance
- POST /transcripts — synchronous public YouTube transcript (1 credit on success)
- POST /transcripts/bulk — async bulk job, up to 100 items
- POST /transcriptions — AI transcription of a media URL
- GET /transcriptions/{id} — retrieve a retained transcription before expiresAt
- GET /jobs/{id} — poll a job; cursor pagination over items
- POST /uploads, POST /uploads/{id}/parts, POST /uploads/{id}/complete, POST /uploads/{id}/resume, GET /uploads/{id}, DELETE /uploads/{id} — multipart local-file upload and AI transcription flow (3 GB limit; cancel before processing refunds reserved credits)

## Product

- [Pricing](https://www.mediacaption.io/pricing): free tier with API access and 10 starter credits; Premium subscription (100 credits/month); one-time credit packs from $5/50 credits
- [FAQ](https://www.mediacaption.io/faq)
- [Terms of Service](https://www.mediacaption.io/terms)
- [Privacy Policy](https://www.mediacaption.io/privacy)

## Cautions for agents

- No idempotency keys: do not auto-retry timed-out POSTs — duplicates spend additional credits. Retry 429s after Retry-After.
- Store transcript content you need longer than 3 days; expired content returns transcription_expired.
- Only job-level webhook events exist in v1; poll GET /jobs/{id} for item detail.