Windows Desktop Touch MCP server

This is a helper that lets your AI assistant actually use your Windows computer. Instead of just chatting, it can look at your screen, find buttons and menus, click them, and type for you. It is handy if you often repeat the same clicks in Notepad, Excel, Chrome, or a terminal and wish the AI could do it.

What this MCP server does

You ask your AI something like open Notepad and type a note. The AI uses this helper to look at your desktop and find the right window and controls. The helper then clicks or types for you, using names of buttons instead of guessing pixel positions. After each action it can check what changed and tell you what happened. You get a short report in the chat, plus screenshots or lists of what is on screen when you need them.

Click to zoom

What you can do with it

• Open apps and switch between windows
• Click buttons and menu items by their name
• Type text into fields, including Japanese, Chinese, and Korean
• Take screenshots of your screen or a single window
• Run commands in a terminal window
• Drag and drop items and scroll through long lists
• Save and reuse SSH or sudo passwords in a secure locker

Try asking your AI

• “Open Notepad and type a short shopping list”
• “Take a screenshot of my Excel window and tell me what is on it”
• “In Chrome, click the search box and type today's weather”
• “Open a terminal and run ipconfig, then show me the result”

What it gives back to you

You get plain answers in the chat: what window is active, what buttons were found, and what changed after a click. It can also send back screenshots or a short list of on-screen elements with their positions. For terminal commands, you see the output text. Everything is described in everyday words, not code.

Before you start

Who it's for

Office workers on Windows who repeat the same clicks and typing every day and want their AI to handle those steps.