Desktop Automation

🚀 Desktop Automation Server

A Model Context Protocol (MCP) server that offers desktop automation capabilities, leveraging RobotJS and screenshot functionality. This server enables Large Language Models (LLMs) to control mouse movements, keyboard inputs, and capture screenshots of the desktop environment.

🚀 Quick Start

Configuration for Using the Desktop Automation Server

Here's how to configure Claude Desktop to use the MCP desktop automation server:

NPX Command

{
  "mcpServers": {
    "desktop-automation": {
      "command": "npx",
      "args": ["-y", "mcp-desktop-automation"]
    }
  }
}

Permission Requirements

This server requires system-level permissions to perform the following actions:

• Capture screenshots
• Control mouse movements and clicks
• Simulate keyboard inputs

When running Claude Desktop for the first time and using this server, you may need to grant these permissions in the security settings of your operating system.

✨ Features

Limitations

Although this server can be used with various MCP clients, it has been primarily tested for compatibility with Claude Desktop.

⚠️ Important Note

The current implementation has a 1MB response size limit. For screenshots, this means:

• High-resolution screenshots may exceed this limit and cause failures.
• Tests have shown that an 800x600 resolution works reliably.
• If you encounter issues, consider reducing the screen resolution or capturing a specific area of the screen.

System Requirements

• Node.js (>=14.x)

💻 Usage Examples

Components

Tools

• get_screen_size

  • Get the screen size.
  • No input parameters required.

• screen_capture

  • Capture the current screen content.
  • No input parameters required.

• keyboard_press

  • Press a key or key combination on the keyboard.
  • Input:
    • key (string, required): The key to press (e.g., "enter", "a", "control").
    • modifiers (array of strings, optional): Modifier keys to hold down while pressing the key. Possible values: "control", "shift", "alt", "command".

• keyboard_type

  • Type text at the current cursor position.
  • Input: text (string, required): The text to type.

• mouse_click

  • Perform a mouse click.
  • Input:
    • button (string, optional, default: "left"): The mouse button to click. Possible values: "left", "right", "middle".
    • double (boolean, optional, default: false): Whether to perform a double click.

• mouse_move

  • Move the mouse to the specified coordinates.
  • Input:
    • x (number, required): The X coordinate.
    • y (number, required): The Y coordinate.

Resource Access

This server provides access to screenshots:

1. Screenshot List (screenshot://list)
   • List the names of all available screenshots.
2. Screenshot Content (screenshot://{id})
   • The PNG image of the screenshot.
   • Accessible via the screenshot ID (named based on the timestamp).