Windows MCP is a lightweight, open-source project designed to integrate AI agents with the Windows operating system. Functioning as an MCP server, it connects large language models (LLMs) with Windows, enabling tasks such as file navigation, application control, and user interface (UI) interaction without requiring specific computer vision techniques. Its rich toolset allows for keyboard and mouse operations, capturing window states, and offers real-time interaction with a latency of 1.5 to 2.3 seconds, depending on system performance. The project is easily customizable, has minimal dependencies, and is available under the MIT license. Users can connect to the MCP server and use various tools like click, type, clipboard, and screenshot functions for effective Windows automation. Caution is advised when deploying in sensitive environments. Contributions are welcome, and setup guidelines are provided for developers wishing to enhance the project. For more details, visit Windows MCP on GitHub.