cursortouch avatar

Windows-MCP

cursortouch/windows-mcp-server

Windows-MCP is an open-source MCP server that enables AI agents to interact with and control the Windows operating system (versions 7-11) through native integration. The server provides tools for file navigation, application control, UI interaction, keyboard and mouse simulation, window state capture, and QA testing, allowing LLMs to perform desktop automation tasks without requiring computer vision or fine-tuned models. It solves the problem of bridging the gap between language models and Windows systems by providing a lightweight, customizable, and LLM-agnostic interface for automating Windows desktop tasks.

GitHub stars
5.6K
tools
21
last updated
Jan 27, 2026

Tools

21 tools

Public tool metadata for what this MCP can expose to an agent.

ClickClick on the screen at the given coordinates.

Click on the screen at the given coordinates.

No parameter schema in public metadata yet.

TypeType text on an element (optionally clears existing text).

Type text on an element (optionally clears existing text).

No parameter schema in public metadata yet.

ScrollScroll vertically or horizontally on the window or specific regions.

Scroll vertically or horizontally on the window or specific regions.

No parameter schema in public metadata yet.

MoveMove mouse pointer or drag (set drag=True) to coordinates.

Move mouse pointer or drag (set drag=True) to coordinates.

No parameter schema in public metadata yet.

ShortcutPress keyboard shortcuts (Ctrl+c, Alt+Tab, etc).

Press keyboard shortcuts (Ctrl+c, Alt+Tab, etc).

No parameter schema in public metadata yet.

WaitPause for a defined duration.

Pause for a defined duration.

No parameter schema in public metadata yet.

WaitForWait until text, an active window, an element, or a focused element appears by polling UI state inside one tool call.

Wait until text, an active window, an element, or a focused element appears by polling UI state inside one tool call.

No parameter schema in public metadata yet.

ControlStatusRead the desktop control owner and remaining user-idle wait.

Read the desktop control owner and remaining user-idle wait.

No parameter schema in public metadata yet.

DisplayInventoryRead display layout, work areas, effective DPI, and scale metadata.

Read display layout, work areas, effective DPI, and scale metadata.

No parameter schema in public metadata yet.

ScreenshotFast screenshot-first desktop capture with cursor position, active/open windows, and an image.

Fast screenshot-first desktop capture with cursor position, active/open windows, and an image.

No parameter schema in public metadata yet.

SnapshotFull desktop state capture for workflows that need interactive element ids, scrollable regions, or use_dom=True browser extraction.

Full desktop state capture for workflows that need interactive element ids, scrollable regions, or use_dom=True browser extraction.

No parameter schema in public metadata yet.

AppLaunch an application by Start Menu name or strictly by executable path with separated argv and optional cwd; resize, move, and switch between windows.

Launch an application by Start Menu name or strictly by executable path with separated argv and optional cwd; resize, move, and switch between windows.

No parameter schema in public metadata yet.

PowerShellTo execute PowerShell commands.

To execute PowerShell commands.

No parameter schema in public metadata yet.

FileSystemRead, write, copy, move, delete, list, search, and inspect files and directories.

Read, write, copy, move, delete, list, search, and inspect files and directories.

No parameter schema in public metadata yet.

ScrapeTo scrape the entire webpage for information.

To scrape the entire webpage for information.

No parameter schema in public metadata yet.

MultiSelectSelect multiple items (files, folders, checkboxes) with optional Ctrl key.

Select multiple items (files, folders, checkboxes) with optional Ctrl key.

No parameter schema in public metadata yet.

MultiEditEnter text into multiple input fields at specified coordinates.

Enter text into multiple input fields at specified coordinates.

No parameter schema in public metadata yet.

ClipboardRead or set Windows clipboard content.

Read or set Windows clipboard content.

No parameter schema in public metadata yet.

ProcessList running processes or terminate them by PID or name.

List running processes or terminate them by PID or name.

No parameter schema in public metadata yet.

NotificationSend a Windows toast notification with title, message, and a required app_id.

Send a Windows toast notification with title, message, and a required app_id.

No parameter schema in public metadata yet.

RegistryRead, write, delete, or list Windows Registry values and keys.

Read, write, delete, or list Windows Registry values and keys.

No parameter schema in public metadata yet.

Sponsors

Readme

Details

Repository
github.com/cursortouch/windows-mcp-server

More from cursortouch