Just say what you want
Describe the pace, style, and destination. The agent uses real editing tools instead of handing back a flattened video you cannot change.
Open source · Conversational video editing
Stop hunting through menus. Tell OpenChatCut what you want and it will organize footage, add captions, lay in music, and build transitions. Every step stays visible and reversible.

It does not just chat. It edits.
It does not merely suggest what to change. It places clips, trims dead space, adds captions, and lays in music. Keep refining the result yourself, or undo any step.
Describe the pace, style, and destination. The agent uses real editing tools instead of handing back a flattened video you cannot change.
Video, audio, captions, transitions, effects, LUTs, motion graphics, and keyframes all live in one project, ready to preview and adjust.
Connect Codex, Claude, or an agent of your own through MCP, plugins, and templates. Change models later without rebuilding your workflow.
The agent knows what you selected and understands the rest of the project. Ask for a faster pace, a different track, or a cleaner transition, and the change appears right on the timeline.

Templates, transitions, effects, audio, and motion graphics are ready to use. Let the agent build a first pass, then refine every detail with familiar editing controls.

One sentence. Then start cutting.
Type it or say it out loud: the story, the pace, the style, and the delivery format.
The agent checks the media, tracks, current selection, transcript, and timeline state.
Each action lands on the timeline with a predictable result, a clear history, and an undo button.
Keep trimming, tuning, swapping media, or export as-is. You are never stuck with a one-shot result.
OPEN SOURCE / OPEN WORKFLOW
OpenChatCut is open source. Pull the code, connect the Codex or Claude setup you already use, and start cutting.