CursorCam plugin
The agent clicks through your web app in a hidden browser and returns an MP4 that zooms in on each click: plugin, MCP and CLI.
- 94 ★ on the repository
- Last commit
Extended access
Why this level
- The agent clicks through your app in a hidden browser and enters login data you give it
- Runs `npx` and writes videos and thumbnails into the project folder
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/cursorcamDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent clicks through your app in a hidden browser and enters login data you give it
- Runs `npx` and writes videos and thumbnails into the project folder
This is third-party code. Review the repository files before installing.
What it does
CursorCam records a demo video of your web app on request to an agent. The agent explores the app, writes the steps and plays them in a hidden browser, then checks stills of the finished video and fixes what looks wrong. The result is a 1080p, 60 fps `video.mp4` and a `poster.jpg` thumbnail in a run folder inside `cursorcam-output/`. The video zooms in where each click happens. If a login or 2FA code is needed the agent asks you, and what you type is used for that run only and never saved. It works as a plugin for Claude Code, Codex, Copilot and Gemini CLI, as the `npx -y cursorcam mcp` MCP server and as a CLI: `npx cursorcam doctor` checks the environment and `npx cursorcam run steps.json` runs a script.
Who it is for. Authors of web products and documentation who need a demo clip of the app with no manual screen recording.
Good fit when
- A release page needs a 20-second clip in which the app signs up a user and creates a project.
- A documentation video is stale after a UI change and must be re-recorded.
- It is convenient to keep the recording script in `steps.json` and rerun it from CI.
- A clip with a close-up on every action is needed so viewers can follow.
Not a fit when
- You need to record a native app or the whole screen: the tool captures only web pages in a browser.
- Node.js 22.13 or Chrome, Edge, Chromium or Brave version 120 or later is missing.
- You need editing with voice and titles: the description covers only recording and zoom.
Example request
Make a 20-second demo of http://localhost:3000: sign up and create a project.Limitations
Node.js 22.13 or newer and one of Chrome, Edge, Chromium or Brave version 120 or later are required: the installed one is used and nothing is downloaded. Video comes out at 1080p and 60 fps. The agent walks through your app and clicks in it, so point it at a local or test environment and keep it away from production data. Login data and 2FA codes are requested from you and used for one run only. The `cursorcam-output` folder should be added to `.gitignore`.
How to disable. Uninstall the plugin in your agent, for example `/plugin uninstall cursorcam@cursorcam`, or remove the MCP server from client config. Delete the `cursorcam-output` folder by hand.
FAQ
What is the output?
A 1080p, 60 fps `video.mp4` and a `poster.jpg` thumbnail in the run folder inside `cursorcam-output/`.
Which agents are supported?
Claude Code, Codex, VS Code with GitHub Copilot, Gemini CLI, Cursor through MCP and Claude Desktop through the `cursorcam.mcpb` bundle.
What about passwords and 2FA codes?
The agent asks you, and what you type is used for that run only and saved nowhere.
Can it run without an agent?
Yes, the CLI: `npx cursorcam doctor` checks the environment and `npx cursorcam run steps.json` plays a script.
Which browser is needed?
Chrome, Edge, Chromium or Brave version 120 or later. The installed browser is used and no new one is downloaded.
Discussion
No questions or reviews yet.
Related
PluginsScreenplay films screen recordings from one prompt: the agent writes a scenario, the tool plays and edits it.
OpenScreen MCP: an agent records the screen, trims and zooms frames, adds captions and exports MP4 or GIF.
The eyes CLI and skill for agents: a video index, search by words and on-screen text, close-up frames and editing metrics
A macOS app with a CLI and skill: record screen and voice, then an agent studies the recording and edits the video.
A self-hosted pipeline turns a text script into an MP4 with voiceover and stock footage, and an agent drives it through MCP.
Claude Video Vision
claude-video-vision
An MCP server and Claude Code plugin that let the agent watch video: frames via ffmpeg and audio transcription via Gemini, local Whisper or OpenAI
README in short
The README lists requirements, an install table per agent, a sample request and how it works. The CLI and MCP have separate guides in `docs/`.