What Is OpenChatCut? An Honest Look at the Open Source Editor
What is OpenChatCut? It is an open source, local-first video editor that lets an AI agent, whether the built-in one or an external one such as Codex or Claude connected through MCP, make real edits to a multitrack timeline instead of just describing what to change. It organizes footage, trims clips, writes captions, and lays in music, and every step stays visible and reversible.
It launched on Product Hunt in July 2026, positioning itself as an open source ChatCut alternative and an AI agent video editor. The sections below cover how it differs from the commercial ChatCut and from one-shot AI generators, and where its rough edges still show, drawn from the project's own documentation and the technical questions its early users asked.
What Is OpenChatCut
OpenChatCut is a video editor built around one idea: an AI agent should be able to act on a real project, not just talk about it. It does not generate a flattened video that can no longer be touched.
Every action the agent takes, placing a clip, trimming dead air, adding a caption, dropping in a transition, gets written to an actual multitrack timeline: video, audio, captions, transitions, effects, LUTs, motion graphics, and keyframes all in the same project. The workflow is straightforward: describe the goal in a sentence, the agent reads the current state of the project, it makes the edit, and the user is free to keep trimming, swap media, or export as is.
The agent doing the work can be the built-in one, or an external agent such as Codex or Claude Code connected through MCP, a protocol that lets outside tools call the editor's functions directly. Users can switch which model drives the editing without rebuilding the project. The project is open source under AGPL and is independent of, not affiliated with, the commercial ChatCut product.
Why It's Showing Up in Searches Right Now
OpenChatCut launched on Product Hunt in July 2026 and picked up momentum quickly: the project has passed 1,300 GitHub stars, per its own site (checked August 26, 2026). Search results for the term right now are dominated by OpenChatCut's own site, its two GitHub mirrors, its Product Hunt page, and a handful of review roundups. Among the eight pages checked for this piece, none laid out what the tool does, how it differs from the commercial ChatCut, and where it is still unfinished, in one place.
When people ask whether a conversational, chat-driven open source video editor exists at all, OpenChatCut usually comes up as the closest match to an editable, reversible timeline, alongside a handful of narrower tools built for shorts, natural-language FFmpeg commands, or generative in-video edits. Anyone comparing that wider field should read the open source ChatCut alternatives roundup; this piece stays with OpenChatCut alone.
OpenChatCut vs a Timeline Editor vs a One-Shot Generator
OpenChatCut sits between two existing categories. A traditional timeline editor gives frame-level precision but no natural-language control. A one-shot AI generator takes a text prompt and hands back a finished video fast, but usually flattened and hard to touch again. OpenChatCut's own comparison, published in its documentation, places itself in between:
| Capability | Traditional timeline editor | One-shot AI generator | OpenChatCut |
|---|---|---|---|
| Track and clip level precision | Yes | No | Yes |
| Modify a project with natural language | No | Yes | Yes |
| Inspectable and undoable changes | Yes | Usually not | Yes |
| Linked transcript and visuals | Partial | No | Yes |
| Direct control from Codex or Claude Code | No | No | Yes |
| Local projects, bring your own key | Depends on the product | Usually cloud only | Yes |
Two distinctions matter more than the table. Against the commercial ChatCut, OpenChatCut describes itself as independent, open source software, not affiliated with that product: it is licensed under AGPL, free to download, and designed to run locally by default. Against one-shot AI generators, its own site makes the contrast explicit: users are never stuck with a one-shot result, because every generated clip, caption, or transition lands on a timeline that stays editable afterward.
What Real Users Asked Before Trusting It
OpenChatCut's Product Hunt launch drew technical questions that the official site and GitHub README do not answer, and the maker, sline, answered each one directly in the comments. Those exchanges are the clearest picture available of how the agent actually behaves once it starts changing a timeline.
Dipankar Sarkar asked the question that matters for any agent holding a stateful project: does it reason from a live state or a stale snapshot, so an instruction like trim the third clip risks hitting the wrong clip once the timeline has moved? The maker's answer: timeline edits are keyed to stable itemId, trackId, and timelineId values rather than ordinal positions, so a lookup fails outright if the referenced item is gone instead of silently editing whatever moved into that slot. One caveat: external MCP calls do not yet carry an expected revision token, so an agent should re-read the timeline after structural edits.
Omri Ben-Shoham asked whether edits made by an external agent through MCP land on a separate draft track that can be rolled back. They do not, yet. External MCP edits currently apply directly to the working timeline, though undo and redo still work. The built-in agent takes a more cautious path: it proposes edits on a scratch document first, lets the user preview and accept or reject them, and detects when a proposal has gone stale before applying it.
Noctis Leonard asked whether the whole editing history is reproducible outside the app, in Git. Not fully. The project's .ccproj.json export bundles the project document, media references, chat, and creative mode, but not the full operation log, pending proposals, or named-version history.
The appeal for other users was simpler. Nishant Dixit, welcoming the launch, said it was timely: after wrestling with AI video tools that paywall you after a couple of clips, a local-first open source option was exactly what was needed.
Where OpenChatCut Still Has Rough Edges
None of this makes OpenChatCut finished software, and the maker has been upfront about that in public. Complete Git-reproducible history, noted above, is one gap. Another is proxy generation for 4K footage: a feature request logged during the launch flagged that scrubbing large footage on a laptop still struggles without it. macOS also shows an unsigned-app warning on the current build, which the maker called expected for now rather than fixed.
OpenChatCut is built for a specific job: an agent editing real footage that already exists, on a real timeline. That is a different problem from generating reusable branded motion graphics from a template. For that second job, AutoAE is a video creation platform where a template is picked from a library, its text, logo, or media placeholders are swapped for a project's own assets, and the result renders in the browser, with no timeline editing involved. One example of the kind of asset that covers, a text animation template used for openers and lower thirds:
The two tools solve different problems and are not substitutes for each other.
FAQ
How is OpenChatCut different from the commercial ChatCut?
OpenChatCut is open source under the AGPL license, free to download, and built to run locally by default, independent of and not affiliated with the commercial ChatCut product. Readers weighing paid alternatives to that commercial tool can see the wider field in the ChatCut alternatives roundup.
Is OpenChatCut really free and able to run locally?
Yes. OpenChatCut is free and open source, and projects and media stay on the user's machine by default rather than in the cloud. One early user summed up the appeal after wrestling with AI video tools that paywall you after a couple of clips, calling a local-first open source option exactly what was needed.
Does the agent reason from a live timeline state, or can position references drift as the project changes?
Timeline edits are keyed to stable itemId, trackId, and timelineId values rather than ordinal positions like the third clip, so a lookup fails outright if the referenced item is gone instead of silently editing whatever moved into that slot. The one gap the maker flagged: external MCP calls do not yet carry an expected revision token, so an agent should re-read the timeline after structural edits.
Do external MCP edits land on a separate draft track that can be rolled back?
Not yet. Edits made by an external agent through MCP currently apply directly to the working timeline rather than a separate draft track, though undo and redo still work. The built-in agent takes a more cautious path, proposing edits on a scratch document first and letting the user preview and accept or reject them.
Can the full editing history be reproduced outside the app, for example in Git?
Not completely. The .ccproj.json export bundles the project document, media references, chat, and creative mode, but not the full operation log, pending proposals, or named-version history. Full Git-reproducible editing history is not implemented yet, according to the maker.