DesktopManager MCP
DesktopManager gives an agent Windows windows and monitors over MCP: it starts read-only and changes need separate flags.
- 108 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- With `--allow-mutations` the agent clicks, types text and launches processes in Windows windows.
- Screenshots and window text can expose personal data.
Install
- Works in
Manual install
Install-Module DesktopManager -Force -VerboseThe first README install command installs the PowerShell module; the `desktopmanager mcp serve` CLI starts the MCP server.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- With `--allow-mutations` the agent clicks, types text and launches processes in Windows windows.
- Screenshots and window text can expose personal data.
This is third-party code. Review the repository files before installing.
What it does
DesktopManager is a Windows desktop automation kit: a .NET library, a PowerShell module, the `desktopmanager` CLI, a tray app and an MCP server. An agent connects with `desktopmanager mcp serve` over stdio, and the server starts in read mode: window and monitor lists, geometry, text and controls of windows, screenshots. Mutating tools such as `click_window_point`, `type_window_text`, `send_window_keys`, `move_window` and `launch_process` turn on with `--allow-mutations`. System settings also need `--allow-system-settings`, and `--allow-process` and `--deny-process` restrict applications, while `--dry-run` previews without acting. Beyond MCP the library manages wallpapers, brightness, resolution, window layouts and audio.
Who it is for. Windows developers and testers who need to give an agent window control and UI checks within clear limits.
Good fit when
- An agent checks whether a window opened and what it shows, without changes.
- A monitor or window screenshot is needed for analysis.
- An autotest clicks application controls and compares a baseline screenshot.
- Window and monitor layouts are saved and restored by script.
Not a fit when
- The system runs Linux: MCP and window adapters target Windows.
- The agent must not be allowed keyboard and mouse input.
- You want to run on a machine with other people's windows without a process allowlist.
Example request
Using DesktopManager, list open windows and capture the editor window without changing anything.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
Windows, tray and MCP work only on Windows; the portable wallpaper and monitor layer also covers macOS, GNOME and KDE. Device and driver management in MCP is read-only. Global airplane mode uses an undocumented COM contract and requires `--allow-experimental`. Clicks and typing into other windows cannot be rolled back, so start with `--dry-run`. Installation comes from the PowerShell Gallery or an MSI release, and MCP needs the ready `desktopmanager` executable.
How to disable. Remove the `desktopmanager mcp serve` entry from the MCP client config. Remove the module with `Uninstall-Module DesktopManager` and the tray app by uninstalling the MSI.
FAQ
How do I start the MCP server?
Run `desktopmanager mcp serve`. It starts in read mode.
How do I allow clicks and typing?
Add `--allow-mutations`. System settings also need `--allow-system-settings`.
How do I limit applications?
The `--allow-process <pattern>` and `--deny-process <pattern>` flags set permitted processes.
Is there an action preview?
Yes, `--dry-run` shows the result of mutating calls without executing them.
What screenshots can it take?
`screenshot_desktop` and `screenshot_window` capture the whole desktop or one window; baselines can be saved and compared.
Discussion
No questions or reviews yet.
Related
MCP serversMCP server and wincli CLI: agents drive Windows apps by control name through UI Automation instead of screen coordinates
An MCP server for driving Windows: the agent opens apps, clicks and types, reads window and UI state, and works with files and PowerShell
The agent moves the mouse, types, drives windows and captures PNG screenshots of a Windows desktop through AutoIt.
An MCP server that lets an agent see the screen and control macOS, Windows and Linux: screenshots, clicks, OCR and screen memory
An MCP server for remote Windows control: screenshots, PowerShell, files, mouse and keyboard, registry and OCR
An MCP server for Windows: the agent moves the mouse, presses keys, manages windows and takes screenshots
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| DesktopManager (this card) | no | not required | 108 | |
| MCP Windows Desktop Automation | no | not required | 119 | Windows window automation over MCP, while DesktopManager starts read-only and adds permission flags and a process filter |
| Windows MCP Server | no | not required | 131 | a Windows control server for agents, while here the same layer is also a PowerShell module, CLI and .NET library |
README in short
The README describes five surfaces: library, module, CLI, MCP and tray, with a selection table. A separate `Docs/DesktopManager.Mcp.md` lists the tools and safety flags.