Desktop Commander MCP Server

Desktop Commander MCP Server

A comprehensive MCP server that gives AI assistants full control over your desktop — monitor system resources, manage windows, capture screenshots, control the clipboard, launch applications, and more.

Category
访问服务器

README

Desktop Commander MCP Server

A comprehensive Model Context Protocol (MCP) server that gives AI assistants like Claude full control over your desktop — monitor system resources, manage windows, capture screenshots, control the clipboard, launch applications, and more.

23 tools across 4 categories, turning Claude into a true desktop assistant.

Why Desktop Commander?

Claude is powerful, but blind and handless on your desktop. Desktop Commander fixes that:

  • "What's eating my CPU?" — instantly see top resource consumers
  • "Take a screenshot" — Claude can actually see your screen
  • "Arrange my windows side by side" — window management by natural language
  • "Copy this to clipboard" — seamless data transfer
  • "Is port 3000 in use?" — debug development issues
  • "Open VS Code in my project folder" — launch apps with context

Tools

System Monitoring (10 tools)

Tool Description
get_system_overview CPU, RAM, disk, swap usage and uptime in one call
list_processes List processes sorted by CPU/memory with optional filtering
get_process_details Deep-dive into a specific process (connections, threads, IO)
get_top_consumers Top N processes by CPU or memory
get_network_stats Network interfaces, bytes sent/received, connection summary
get_cpu_per_core Per-core CPU usage and frequency
kill_process Gracefully terminate or force-kill a process (with safety guards)
get_battery_status Battery level, charging status, time remaining
find_process_by_port Find which process is using a specific port
get_temperatures Hardware temperature sensors

Window Management (5 tools)

Tool Description
list_windows List all visible windows with title, position, size, PID, and state
focus_window Bring a window to the foreground by title or handle
arrange_window Move and/or resize a window to exact pixel coordinates
set_window_state Minimize, maximize, or restore a window
close_window Gracefully close a window

Desktop Control (5 tools)

Tool Description
capture_screenshot Screenshot full screen or a specific window (returns file path or base64)
read_clipboard Read text content from the system clipboard
write_clipboard Write text to the system clipboard
send_notification Show a desktop toast notification
get_display_info Monitor info: resolution, DPI, scaling, multi-monitor layout

Application Management (3 tools)

Tool Description
launch_application Start a program, open a file, or open a URL
search_installed_apps Search installed software (from Windows registry)
get_startup_programs List programs configured to run at startup

Installation

git clone https://github.com/19miha99/system-monitor-mcp.git
cd system-monitor-mcp
pip install -e .

Usage with Claude Desktop

Add to your Claude Desktop config (claude_desktop_config.json):

{
  "mcpServers": {
    "desktop-commander": {
      "command": "python",
      "args": ["-m", "system_monitor_mcp.server"]
    }
  }
}

Usage with Claude Code

claude mcp add desktop-commander -- python -m system_monitor_mcp.server

Example Workflows

System Health Check

"Give me a complete system health check"

Claude calls get_system_overview, get_top_consumers, get_network_stats and gives you a comprehensive report.

Debug a Slow Computer

"My computer is slow, what's going on?"

Claude checks get_top_consumers for CPU and memory, identifies the culprit, and offers to kill_process if needed.

Window Arrangement

"Put Chrome on the left half and VS Code on the right half of my screen"

Claude uses list_windows to find them, get_display_info to know the screen size, then arrange_window to position each one.

Port Conflict Resolution

"Something is using port 8080, find and kill it"

Claude calls find_process_by_port(8080), shows you what's there, and kill_process to free the port.

Visual Debugging

"Take a screenshot and tell me what you see"

Claude uses capture_screenshot, then reads the image to analyze what's on screen.

App Launcher

"Open my project in VS Code"

Claude calls launch_application("code", args=["C:/Users/you/project"]).

Architecture

src/system_monitor_mcp/
├── app.py          # FastMCP instance (shared across modules)
├── server.py       # Entry point — imports all tool modules
├── helpers.py      # Shared utilities (format_bytes, format_uptime)
├── monitor.py      # 10 system monitoring tools
├── windows.py      # 5 window management tools
├── desktop.py      # 5 desktop control tools (screenshots, clipboard, etc.)
└── apps.py         # 3 application management tools

Requirements

  • Python 3.10+
  • psutil — cross-platform system monitoring
  • mcp — Model Context Protocol SDK
  • Pillow — screenshot capture
  • pywin32 — Windows API access (auto-installed on Windows)

Platform Support

Category Windows Linux macOS
System Monitoring Full Full Full
Window Management Full Planned Planned
Screenshots Full Planned Planned
Clipboard Full Planned Planned
Notifications Full Planned Planned
App Management Full Planned Planned

Primary platform: Windows. Cross-platform support for desktop features is planned.

Safety

  • System-critical processes (PID 0, 1, 4) cannot be killed
  • The server refuses to kill its own process
  • Windows are closed gracefully via WM_CLOSE (not force-terminated)
  • All operations provide clear error messages on failure

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选