OpenChatCut
A free, open-source video editor you can talk to: describe the cut you want in plain English and an AI agent makes the edit on a real timeline you can still fine-tune by hand.
🔗 Visit OpenChatCut
Description
Most video editors ask you to learn a timeline, dozens of keyboard shortcuts, and where every button lives before you can trim a single clip. OpenChatCut flips that: you type or say what you want — "cut the dead air at the start" or "add captions and speed up the boring part" — and an AI agent makes the edit for you, the same way you'd direct a human editor sitting next to you.
OpenChatCut is a free, open-source (AGPL-3.0+) desktop app for macOS, Windows, and Linux that pairs a conventional multitrack timeline with an AI agent layer driven through the Model Context Protocol (MCP), so it can be pointed at Claude, Codex, or a custom model. Every AI-driven action — cuts, transitions, captions, LUTs, motion graphics, keyframes — lands on a real, undoable timeline rather than a black box, so you can always drop back to manual editing. It runs local-first (your project and media stay on your machine, with optional Cloudflare R2 storage), supports hardware acceleration (VideoToolbox on macOS, NVENC on Windows), and exposes its own local MCP endpoint so other tools can drive it programmatically. It positions itself explicitly as an open-source alternative to the commercial ChatCut.
💬 Our review
The short version: if you like the idea of ChatCut — editing video by chatting with an AI — but don't want to pay a subscription or hand your footage to a cloud service, OpenChatCut is the open-source version of that same idea, and it actually works on a real, editable timeline rather than just spitting out a rendered result.
Against ChatCut itself, OpenChatCut trades a more polished managed experience and multi-platform cloud sync for zero cost, full local control of your media, and the ability to point it at whichever AI model you prefer (Claude, Codex, others) via MCP instead of being locked to one vendor. Against traditional editors like Premiere or DaVinci Resolve, it's not trying to compete on manual editing depth — it's for people who'd rather describe an edit than click through menus for it, with the safety net of falling back to a normal timeline when the AI gets it wrong. At 500+ GitHub stars it's a young but active project, so expect occasional rough edges rather than the stability of a decade-old NLE; it's a strong pick for creators comfortable with open-source software who want AI-assisted editing without a subscription, less so for teams who need guaranteed support SLAs.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Licence AGPL-3.0+, aucun abonnement ni frais caché, auto-hébergé (installeurs macOS, Windows, Linux)
Pros
Gratuit et open-source (AGPL-3.0+), pas d'abonnement
Montage IA sur timeline réelle et annulable, pas une boîte noire
Choix du modèle IA (Claude, Codex, custom via MCP)
Local-first : projets et médias restent sur la machine
Cons
Projet jeune (~500 étoiles GitHub), moins mature que les éditeurs établis
Pas de support commercial garanti
Nécessite du matériel correct pour l'accélération vidéo
Installation depuis les sources demande Node.js 24.x