npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@one710/computermate

v0.3.0

Published

MCP server exposing computer-use tools to LLMs

Readme

🤖 ComputerMate

"Your AI's hands and eyes on any machine."
(The forbidden MCP of the AI era...)

Test Status npm version npm downloads License: MIT Platform

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 native mode.
  • 🐳 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.