69 built-in skills
Type / to invoke skills covering voiceovers, social, color, motion, audio and workflows — or create your own.
Bring your footage and describe the result you want. The agent adds captions, music and transitions, then you refine picture and sound on the timeline until the cut feels right.
Works out of the box · local models · 7-day free trial
Free · AGPL open source · bring your own models

By the numbers
01Agent capabilities
Converse, cut, refine: every agent step is undoable.
Cut this 20-minute talking-head into 3 highlight shorts
Done: 3 vertical shorts are on the timeline, ready to refine.
More things you can say
Type / to invoke skills covering voiceovers, social, color, motion, audio and workflows — or create your own.
Save a cut you like as a reference project and cite it with # to apply the same treatment to new footage.
Agent edits share the same command path as manual edits — roll back anytime and keep adjusting by hand.
Bring titles, captions and music into one project. Listen back, balance the soundtrack and refine the transitions between clips.
Try a direction like this
Add a “Seaside Diary” title and a gentle soundtrack to this beach video. Keep the sound of the waves.

02Extend as you go
Whatever's missing, one line away.
03Model freedom
No vendor lock-in — 20+ model providers and the mainstream media services, plug in with your own key.
Supported LLM providers
Media generation & open stock
Local CLI agents
Signed-in CLIs work as the agent model with no extra keys.
Models and media services are billed by their providers — never bundled into the license.
Pro's local speech recognition, matting and denoising run on your machine once downloaded.
Let Claude Code, Codex or Cursor drive your project over MCP.
04Two tools, one app
Say it; the agent edits the timeline.
Multitrack timeline, every frame in your control.
One project · every step undoable

05Local first
Editing never depends on the cloud — it keeps working offline.
Timelines and media files live on your disk — copy or back them up anytime.
Conversations stay local; nothing is uploaded to OpenChatCut.
Keys are stored in the local database and never uploaded.
Speech recognition, matting and denoise run offline once downloaded.
When cloud models or online asset services are enabled, only what the task requires is sent to the provider you chose.
06Guides & use cases
Hands-on guides: transcripts, long-to-short cuts, captions, skills.
07Editions
The two editions are separate pieces of software for different ways of working.
The source is public on GitHub, and you can use and modify it under the AGPL. A good fit for developers and creators who want to study the code, adapt it or build it themselves.
A proprietary desktop editor built for performance: Tauri native shell, Rust rendering core and offline local models. Installers from this site, for creators who want recording and two-way MCP out of the box.
08FAQ
Understand the open-source edition, Pro and AI model services before choosing your workflow.
OpenChatCut
Download the open-source edition and try an edit with the agent, or explore the Pro desktop workflow.