OpenScreen — Claude Code MCP server for screen recording
OpenScreen records your screen and lets an agent turn a take into a product demo. Use it to show an app change or attach a video to a pull request.
Open on GitHubWhat OpenScreen does
OpenScreen can capture a window or the whole screen, then trim the recording, add zooms, captions, and annotations. An agent can plan a storyboard, record the app, edit the take, export an MP4, and attach it to a pull request. The CLI also supports headless recording and export, and lets scripts edit the .openscreen project JSON.
AI-assisted editing needs your own key from Claude, OpenAI, Gemini, Mistral, OpenRouter, MiniMax, or an OpenAI-compatible endpoint. The assistant is off by default, so set up a key before asking it to edit. Voiceover captions are transcribed on-device and work offline. Exports include MP4 and GIF, with CPU fallback if GPU rendering fails. The project is under active development; updates can break the .openscreen format or CLI.
What it helps with
- record an app walkthrough for a product demo
- trim a recording using its transcript
- add zooms and annotations to a screen recording
- export a recording as MP4 or GIF
- attach a demo video to a pull request
Example prompts for the agent
- Record the app and make a demo video for this pull request.
- Trim the recording using its transcript, add zooms, and export an MP4.
- Turn this screen recording into a GIF.
How to install OpenScreen
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about OpenScreen
How do I install OpenScreen in Claude Code?
Add the MCP server with claude mcp add openscreen -- "<executable>" mcp, then install the demo skill with npx skills add getopenscreen/openscreen --skill openscreen-demo. Replace <executable> with the OpenScreen app executable on your system.
What does OpenScreen do?
It records your screen and gives an agent editing tools for trimming, zooming, annotations, and video export. The agent can make a demo of an app change and attach it to a pull request.
Does OpenScreen need an API key?
AI-assisted editing requires your own key from a supported provider or an OpenAI-compatible endpoint. The assistant is off by default; voiceover transcription runs on-device and works offline.
