mcp-windows-app-launcher

mcp-windows-app-launcher

Enables AI assistants to find and launch Windows applications by name using multi-strategy discovery via registry, Start Menu, PATH, and common directories.

Category
访问服务器

README

mcp-windows-app-launcher

中文文档

🚀 A Model Context Protocol (MCP) tool that enables AI assistants like Claude to instantly find and launch Windows applications by name.

Project Overview

mcp-windows-app-launcher is an MCP Server that exposes a single launch_windows_app tool. When an AI assistant invokes it, the server searches your Windows system through multiple strategies — registry, Start Menu, PATH, and common install directories — to locate the target application's .exe path and launch it immediately.

Built with the official MCP Python SDK, it integrates seamlessly with Claude Desktop and Claude Code via stdio transport.

Features

  • Multi-strategy app discovery — searches registry App Paths, Start Menu .lnk shortcuts, system PATH, and Program Files directories in priority order
  • Smart name aliasing — maps common names to their real executables (e.g. vscodecode, edgemsedge, terminalwt)
  • Async process launching — uses asyncio.create_subprocess_exec for non-blocking app startup
  • Robust error handling — clear, Chinese-localized error messages for missing apps, permission issues, and system errors
  • Zero external dependencies beyond MCP SDK — all Windows interaction uses stdlib (winreg, subprocess, asyncio, pathlib)

Installation & Setup

Prerequisites

  • Windows 10 or later
  • Python 3.10+

1. Clone the repository

git clone https://github.com/ye1430937615-ai/mcp-windows-app-launcher.git
cd mcp-windows-app-launcher

2. Create a virtual environment and install dependencies

python -m venv .venv
.\.venv\Scripts\python -m pip install -r requirements.txt

3. Verify the installation

.\.venv\Scripts\python -c "from app_locator import find_app_path; print(find_app_path('notepad'))"

You should see the absolute path to notepad.exe printed on your terminal.

How to Configure

Configure your MCP client to launch this server. Choose the section that matches your client.

Claude Desktop

Add the following to your Claude Desktop mcp.json configuration file:

Windows path%APPDATA%\Claude\mcp.json

{
  "mcpServers": {
    "mcp-windows-app-launcher": {
      "command": "C:\\Users\\ye1430937615-ai\\path\\to\\mcp-windows-app-launcher\\.venv\\Scripts\\python.exe",
      "args": [
        "C:\\Users\\ye1430937615-ai\\path\\to\\mcp-windows-app-launcher\\server.py"
      ]
    }
  }
}

[!WARNING] You MUST replace all placeholder paths above with the actual absolute paths on your machine.

  • All backslashes in the JSON must be escaped as double backslashes (\\).
  • The Python executable must be the one inside .venv\Scripts\python.exe, not the system Python.

Tip: Run the included helper script to generate the correct JSON for your machine automatically:

.\.venv\Scripts\python generate_config.py

Claude Code (CLI)

Add the following to your Claude Code settings file:

Windows path%USERPROFILE%\.claude\settings.json

{
  "mcpServers": {
    "mcp-windows-app-launcher": {
      "type": "stdio",
      "command": "C:\\Users\\ye1430937615-ai\\path\\to\\mcp-windows-app-launcher\\.venv\\Scripts\\python.exe",
      "args": [
        "C:\\Users\\ye1430937615-ai\\path\\to\\mcp-windows-app-launcher\\server.py"
      ]
    }
  }
}

Or use .claudecode/mcp.json in your project directory for per-project configuration:

{
  "mcpServers": {
    "mcp-windows-app-launcher": {
      "type": "stdio",
      "command": ".\\.venv\\Scripts\\python.exe",
      "args": ["server.py"]
    }
  }
}

4. Restart your MCP client

Restart Claude Desktop or Claude Code. The launch_windows_app tool should now be available.

Send a message like:

Please launch Notepad for me

The assistant will call the tool with app_name: "notepad" and the application will start.

Usage Examples

Prompt Tool Call
"Launch Firefox" launch_windows_app("firefox")
"Open VS Code" launch_windows_app("code")
"Start Windows Terminal" launch_windows_app("terminal")
"Open Calculator" launch_windows_app("calc")
"Launch Task Manager" launch_windows_app("taskmgr")

How It Works

User prompt → Claude → MCP Tool call → server.py → app_locator.py
                                                       │
                                          ┌────────────┼────────────┐
                                          ▼            ▼            ▼
                                     Registry      Start Menu    PATH / Dirs
                                          │            │            │
                                          └────────────┼────────────┘
                                                       ▼
                                               Found .exe path
                                                       │
                                                       ▼
                                          asyncio.create_subprocess_exec()
                                                       │
                                                       ▼
                                                  App launches 🎉
  1. server.py receives the launch_windows_app tool call from the MCP client
  2. app_locator.find_app_path() searches in priority order:
    • Registry App Paths (HKLM x64, HKLM x32, HKCU) — the standard Windows app registration
    • Start Menu .lnk shortcuts — resolved via PowerShell COM
    • System PATH — via shutil.which()
    • WindowsApps — Win10/Win11 App Execution Aliases
    • Program Files directory scan — depth-limited to 3 levels
  3. app_locator.launch_app() starts the found executable via asyncio.create_subprocess_exec()

License

MIT — see LICENSE for details.

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