The analysis tool is free and reports what a job will cost before it runs, which is the right first call on a metered pipeline. Clip listings come back rating-sorted and stripped of long signed URLs by default — a small decision that keeps a hundred-clip result readable. Because each tool carries a read, write or destructive annotation, a client can tell which calls to hold for approval.
A server over the CutPro API covering the whole clipping pipeline: upload or point at a video, run AI clipping, review the clips, apply an editing template, render to MP4, and publish to connected social accounts immediately or on a schedule.
- A video analysed for metadata and credit cost before you commit anything, and your source library listed
- Your own file uploaded through a presigned URL and registered when the upload completes
- Clipping submitted and polled to completion, with submissions listed per video
- Clips of a finished submission listed in rating order, with long signed URLs omitted unless you ask for them
- An editing template applied across clips in bulk
- Clips rendered to MP4, polled, downloaded individually or bundled together, with the workspace's render quota readable
- Posts created, scheduled, published, updated and retried per target platform
- Connected social accounts listed
- The workspace this key resolved to, its credit balance, and the ledger behind it
- Read-only, write and destructive annotations on every tool so a client can plan which calls need a human
A CutPro account on the paid tier and an API key generated in your account settings. Node for the local path, or the hosted endpoint with a browser sign-in if you would rather install nothing.
One command plus a key — npx -y @cutpro/mcp, then supply credentials