MCPComputerUse

MCPComputerUse is an MCP server that provides native desktop automation on Windows 10 and 11 via direct Windows API integration. Built in C# on .NET 8.0, it connects AI clients directly to system-level OS capabilities without requiring external runtimes or complex Python dependencies. Desktop power users, software testers, and AI engineers use this tool to allow LLMs to perceive screens and control applications. Through its standardized Model Context Protocol tools, it enables clients to capture multi-monitor or window-specific screenshots, enumerate active application windows, manage window states like minimizing and maximizing, simulate precise mouse movements and clicks, type Unicode text, and trigger modifier keys. Additionally, the server supports sequential macro execution to chain multiple UI actions together, providing high-performance desktop control with minimal memory overhead and fast execution response times.

Category: Browser & Web Automation

Tags: automation, csharp, desktop, win32, windows

Visit MCPComputerUse

How to install and configure MCPComputerUse

Follow these steps to set up MCPComputerUse with Claude Desktop: 1. Download the latest release executable (MCPComputerUse.exe) from the project repository or build it using dotnet publish src/MCPComputerUse -c Release -r win-x64 --self-contained -o ./publish. 2. Extract or move MCPComputerUse.exe to your chosen local directory. 3. Open your Claude Desktop configuration file (claude_desktop_config.json). 4. Add the server under the mcpServers configuration key: json { "mcpServers": { "computer-use": { "command": "c:/path/to/MCPComputerUse.exe", "args": [] } } } 5. Replace c:/path/to/MCPComputerUse.exe with your actual file path and restart Claude Desktop.

What you can do with MCPComputerUse

• Capturing desktop or window-specific screenshots across multi-monitor setups to analyze user interfaces during automated testing. * Automating form filling and repetitive data entry by sending mouse clicks and typing Unicode text sequences directly into Windows applications. * Managing open software windows by querying active window lists and programmatically focusing, minimizing, maximizing, or restoring windows. * Executing automated testing macros that combine window focusing, cursor navigation, button clicks, and keystrokes into timed sequences.

Key facts

• Open Source
• https://github.com/kblood/MCPComputerUse
• Browser & Web Automation, Cloud & Infrastructure
• automation, csharp, desktop, win32, windows

Part of MCP Servers

What is MCPComputerUse?

MCPComputerUse is an open-source C# implementation of an MCP server for Windows automation. It uses native Windows API calls via P/Invoke to provide screenshot capture, window management, and input automation capabilities to AI assistants.

How do I install MCPComputerUse?

Download the prebuilt MCPComputerUse.exe file from the GitHub repository release page, place it in a local directory, and configure your MCP client to launch the executable. Because it is published as a self-contained executable, no separate .NET runtime installation is required.

Which MCP clients work with MCPComputerUse?

MCPComputerUse works with Claude Desktop and any standard Model Context Protocol client capable of spawning local process servers over standard input and output on Windows 10 or Windows 11.

Is MCPComputerUse open source?

Yes, MCPComputerUse is open source and released under the MIT License. The complete source code, including the native automation and protocol libraries, is accessible on GitHub for modification and contribution.

Does MCPComputerUse require administrator privileges?

No administrator privileges are required for standard desktop operations. However, some target applications protected by User Account Control or administrative security policies may require the server to run elevated to accept simulated input or window focus.