AI video editing platform
The machine finds the cut. You still call it.
Cutroom watches your footage end to end — scenes, speech, faces, motion — and hands back an edit plan you can interrogate frame by frame before a single pixel renders.
The journey
Six passes between raw footage and a finished cut.
Each pass is visible, inspectable and reversible. The system proposes; you stay the editor of record.
- 01
Ingest
Drop an mp4, mov or mkv. It streams straight to object storage with real progress — no waiting room.
- 02
Scene detection
Every cut, dissolve and hold is mapped into precise scene bounds you can trim against.
- 03
Transcription
Word-level transcript, aligned to the timeline, clickable all the way down to the syllable.
- 04
Vision tagging
Faces, motion, framing and on-screen events become searchable markers on the track.
- 05
Edit plan
A reviewable plan of keeps and cuts with captions and overlays — proposed, never forced.
- 06
Render
Approve the plan and the cut renders out, tracked job by job, ready to download.
The review room
A timeline you can argue with.
Keeps and cuts are colour-coded and labelled with an icon and text, so the decision never depends on colour alone. Drag trim handles inside scene bounds, edit captions inline, then save the batch in one explicit action.
- ✓ Keep segment
- ✕ Cut segment
- ◆ Event marker
Built like production software, not a demo reel.
Plans, not black boxes
Every proposed cut arrives as a versioned, inspectable edit plan. Revise it in plain language; history is never overwritten.
Pipeline you can watch
Scene detection, transcription, timeline extraction, vision tagging and plan generation each report their own state, live.
Single sign-on, properly
Hosted identity with MFA support, tokens held in httpOnly cookies, and silent refresh long before anything expires.
Bring the footage. Keep the final say.
Sign in with your organisation account to upload your first video.