ComputerMate
Enables AI to control a computer through mouse, keyboard, and screen capture tools, with support for local native and Docker sandboxed environments.
README
🤖 ComputerMate
"Your AI's hands and eyes on any machine."
(The forbidden MCP of the AI era...)
ComputerMate is a powerful Model Context Protocol (MCP) server that grants AI models the ability to interact with your computer just like a human would. From taking screenshots to clicking buttons and typing text, ComputerMate exposes your local machine (or a sandboxed Docker container) as a set of tools for LLMs.
✨ Features & Highlights
- 🖥️ Cross-Platform: Unified support for Linux, macOS, and Windows via a single
nativemode. - 🐳 Docker-First: Run in a fully isolated Linux desktop environment with VNC access and state persistence.
- 🖱️ Virtual Cursor: Enable a visual mouse cursor in Playwright by setting
VIRTUAL_CURSOR=true. - 🌐 Web Automation: Integrated Playwright support for high-performance, browser-only computer use.
- 📐 Smart Scaling: Automatically scale screenshots and coordinates to fit LLM context limits via
MAX_SCALING_DIMENSION.
🛠️ Available Tools
ComputerMate exposes the following tools to the LLM:
| Tool | AI-Friendly Description |
|---|---|
screenshot |
Take a full screenshot of the current screen or browser viewport. |
screenshot_region |
Capture a specific rectangular area by providing two diagonal points. |
click |
Move pointer and click (left, middle, right supported). |
double_click |
Rapidly click twice at the given coordinates. |
scroll |
Scroll the window content at (x, y) by given amount. |
type |
Send keyboard text input to the active window. |
keypress |
Send key combinations (e.g. ["ctrl", "c"], ["alt", "tab"]). |
move |
Move the mouse pointer without clicking. |
drag |
Drag the mouse from start point along a path of coordinates. |
wait |
Pause execution for a set number of milliseconds. |
get_dimensions |
Retrieve the screen or viewport width and height. |
get_environment |
Returns the current platform (linux, macos, windows, browser). |
goto |
(Playwright only) Navigate to a specific URL. |
back |
(Playwright only) Go back in history. |
forward |
(Playwright only) Go forward in history. |
get_current_url |
(Playwright only) Retrieve the current active page URL. |
🚀 Quick Start
📦 Using npx (Local)
Run the server directly without installing:
# For local OS interaction (macOS, Windows, Linux)
npx @one710/computermate native
# For Playwright (Browser only)
npx @one710/computermate playwright
🐳 Using Docker (Safe & Persistent)
The recommended way to use ComputerMate is via Docker. This provides a sandboxed environment and VNC access.
# Clone the repository
git clone https://github.com/one710/computermate.git
cd computermate
# Start the server with persistence
docker-compose up --build
- MCP Endpoint:
http://localhost:3000 - VNC View:
localhost:5900(Password:one710)
🛠️ Prerequisites & Installation
🐧 Linux (Ubuntu / Debian Desktop)
sudo apt-get update
sudo apt-get install -y libxtst-dev libpng-dev xvfb xdotool
🍎 macOS
Ensure you have granted Accessibility permissions to your terminal or IDE (e.g., Cursor, VS Code, iTerm2) in System Settings > Privacy & Security > Accessibility.
🪟 Windows
No external binaries are required!
⚙️ Configuration
Environment Variables
| Variable | Purpose | Default |
|---|---|---|
COMPUTER_TYPE |
(Docker / HTTP) Sets the computer backend: native or playwright. |
native |
MAX_SCALING_DIMENSION |
Caps the max width or height of screenshots (e.g., 1024x768). Scales coordinates. |
None |
VIRTUAL_CURSOR |
(Playwright) Shows a visual red dot where the "mouse" is. | false |
HEADLESS |
(Playwright) Runs the browser in headless mode. | false |
🏗️ Local Development
To run from source:
# 1. Install dependencies
npm install
# 2. Build the project
npm run build
# 3. Run the server
node dist/index.js native
⚠️ Important Warnings
[!CAUTION] This tool grants significant control over your system. An LLM with access to ComputerMate can read your files, access your browser, and execute commands as the current user.
- Use with Caution: Do not run this on a machine with sensitive data or production access.
- Sandbox Recommended: Use the Docker environment for maximum safety.
- Monitor closely: Always keep an eye on what the AI is doing.
🛠️ MCP Configuration
Add this to your claude_desktop_config.json (or equivalent MCP client config):
{
"mcpServers": {
"computermate": {
"command": "npx",
"args": ["-y", "@one710/computermate", "native"],
"env": {
"MAX_SCALING_DIMENSION": "1024x768"
}
}
}
}
🌉 Stdio-to-Docker Bridge (mcp-remote)
If your MCP client (like Cursor or older versions of Claude Desktop) only supports stdio transports, you can bridge it to the ComputerMate Docker container:
{
"mcpServers": {
"computermate": {
"command": "npx",
"args": [
"-y",
"mcp-remote@latest",
"http://localhost:3000/mcp",
"--transport",
"http-only"
]
}
}
}
This command starts a local stdio server that transparently forwards all requests to the ComputerMate container running at http://localhost:3000.
📄 License
This project is licensed under the MIT License. See the LICENSE file for details.
<p align="center"> Built with ❤️ by the <b>One710 Softworks</b> Team. </p>
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.