Public API + MCP server, agent-ready
Clip from your own code or an AI agent: a REST API (submit, poll, list, usage, transcript) and an MCP server with sixteen tools for Claude, Cursor and ChatGPT. Safe retries with Idempotency-Key, scoped read-only keys, signed webhooks and a published OpenAPI 3.1 spec.
A REST API and an MCP server over the same pipeline the web app uses. Submit a video URL, poll the job, read the clips, the transcript and your usage.
Three things in it are unusual enough to name. Idempotency keys: retrying a submission with the same key returns the original job instead of creating a second one, which matters when a network blip makes your client retry a video you already paid for. Signed webhooks: the payload carries an HMAC over a timestamp and the body, so you can tell our call from anyone else's. Scoped keys: a read only key cannot start a job, so the credential in your analytics script cannot spend your quota.
The quota is shared across every surface. A video clipped from the API counts the same as one clipped from the dashboard, and there is no per minute surcharge on top.
The OpenAPI 3.1 spec is published, so a client generator can read it rather than you reading our docs.
2 videos a month, no card. What is still being built is on the roadmap.