computer-control-mcp
Computer Control MCP is a server designed for manipulating various computer control tasks like mouse movements, keyboard typing, and screen management using technologies such as PyAutoGUI and RapidOCR. It operates without external dependencies and provides functionalities similar to Anthropic's 'computer-use', optimized for Windows.
Computer Control MCP
Overview
MCP server that provides computer control capabilities like mouse, keyboard, OCR, etc. using PyAutoGUI, RapidOCR, and ONNXRuntime. It's similar to 'computer-use' by Anthropic. The server operates without external dependencies and has only been tested on Windows but is expected to work on other platforms as well.
Features
- Control mouse movements and clicks
- Type text at the current cursor position
- Capture screenshots with OCR
- Manage windows (list and activate)
- Keyboard actions
- Drag and drop operations
Usage
Install using uvx
or pip
to set up the server. Subsequent runs are instant once dependencies are downloaded.
Development
Clone the repository, set up the environment, and start the server with specified commands. Tests can be run using pytest
.
License
MIT