Real timeline editing
OpenChatCut uses an actual multitrack project, so AI actions land on clips, captions, transitions, effects, LUTs, motion graphics, and keyframes rather than producing a flattened output you cannot change.
OpenChatCut
Freemium
OpenChatCut — бесплатный видеоредактор с открытым исходным кодом и локальным приоритетом для диалогового монтажа с помощью AI на настоящей многодорожечной временной шкале. Превращайте промпты в редактируемые фрагменты и сохраняйте контроль над проектами, экспортом и изменениями таймлайна.
Что такое OpenChatCut?
OpenChatCut is an open-source, local-first video editor built around AI-assisted editing on a real multitrack timeline. Instead of asking a model to generate a finished clip, it lets you describe the goal of an edit and then applies the changes directly to the project, where they remain visible, adjustable, and undoable.
The product is aimed at people who want conversational editing without giving up timeline control. The site positions it as a fit for interviews, talking-head videos, podcasts, and product demos, and it supports workflows where creators or teams can refine an AI-assisted first pass with normal editing tools before export.
Что умеет OpenChatCut?
Conversational agent workflow
You can describe an edit in natural language, and the agent applies changes directly to the project state. The workflow supports iterative refinement with visible history and undo.
Model and agent flexibility
The built-in agent supports multiple providers and OpenAI-compatible APIs, and external agents such as Codex and Claude Code can work on the same project through MCP.
Local-first project storage
Projects and media stay local by default, which is important for large or private footage. The about page says online models, media search, and remote storage are optional add-ons.
Editable export options
The docs say exports can include video, audio, captions, and project formats such as FCPXML, depending on the current release.
External tool integration
The docs describe a local Streamable HTTP MCP endpoint for connecting external tools to the editor, with a default endpoint on localhost.
Варианты использования
“Draft an edit from a prompt”
Use OpenChatCut to turn a rough script or spoken request into a first-pass edit with cuts, captions, music, and transitions, then refine the result on the timeline.
“Edit long-form spoken content”
Use it for interview, talking-head, podcast, or product-demo footage where you need to adjust pacing, remove dead space, and keep control over the final cut.
“Work with external agents”
Use the editor when you want to connect Codex, Claude, or another MCP-capable agent to a real project rather than a black-box generator.
“Keep projects local”
Use the local-first setup when footage is private or large and you want media and project data to stay on your machine by default.
“Prepare handoff-friendly exports”
Use the export workflow when you need deliverables that remain editable or reviewable, such as captions or project files, instead of only a rendered video.
Часто задаваемые вопросы
Is OpenChatCut free?
OpenChatCut is free to download and use. The pricing page says it does not sell seats or credit packs on the site, and the product is available as open source.
Where do projects and media live?
Projects and media stay on your machine by default. The about page says the default location is ~/.openchatcut, with IndexedDB used as a browser cache, and that cloud services are optional rather than required.
How is this different from AI video generators?
OpenChatCut edits a real multitrack timeline instead of producing a flattened one-shot video. That makes the result previewable, tweakable, and undoable, and it keeps tracks, captions, transitions, effects, and history in the project.
Which AI models or agents does it support?
The docs say the built-in agent supports multiple providers and OpenAI-compatible APIs, and that external agents such as Codex and Claude Code can drive the project through MCP.
What is MCP used for?
OpenChatCut exposes a local Streamable HTTP MCP endpoint for connecting external agents. The docs list the default endpoint as http://localhost:5199/api/external-mcp/mcp.
Why does macOS warn that the app is unsigned?
The docs note that early desktop builds may be unsigned, so macOS can warn on first launch. Users may need to approve the app in system settings.
What can I export?
The docs say you can export video, audio, captions, and project formats such as FCPXML, with capabilities depending on the current release.
Краткие сведения
- Category
- Video Editing / Developer Tool
- Model
- Free and open source
- Workflow
- Conversational editing on a multitrack timeline
- Storage
- Local-first by default
- Integrations
- Built-in providers, OpenAI-compatible APIs, MCP, Codex, Claude Code
- Source
- openchatcut.com
Анализ трафика OpenChatCut
Данные о трафике приведены только для справки.
- Рейтинг домена
- 10