Skip to main content
Glama
Sign In
deenesjakoruzh
UIAutomationMcp

UIAutomationMcp

by trsdn
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Local

UIAutomationMcp

License: MIT .NET Platform

Windows UI 自动化工具包，提供三个对齐的入口点：

  • CLI（uiamcp）

  • MCP 服务器（UIAutomationMcp.McpServer）

  • VS Code 扩展（uia-mcp）

安装

VS Code 扩展（推荐）——在 Marketplace 中搜索“UI Automation MCP”，或从最新版本安装 UIAutomationMcp-<version>.vsix。自包含；捆绑了 MCP 服务器、CLI 和两个代理技能。

Claude Desktop —— 从最新版本下载 uia-mcp-<version>.mcpb 并双击它。

NuGet .NET 工具

dotnet tool install --global UIAutomationMcp.McpServer   # MCP server -> uiamcp-server
dotnet tool install --global UIAutomationMcp.CLI         # CLI        -> uiamcp

独立 ZIP —— UIAutomationMcp-MCP-Server-<version>-windows.zip 包含自包含的 MCP 服务器和 CLI；无需 .NET 运行时。

需要 Windows 和交互式桌面会话。UI 自动化无法在无头模式下运行。

Related MCP server: windows2text

谱系

此仓库是更广泛的 MCP 自动化谱系的一部分，并有意保留对 Stefan Brönner 早期工作的归属。

它还保留了对 Excel MCP 的历史引用，作为该谱系中的相关兄弟项目，尽管此仓库本身专注于 Windows UI 自动化。

功能

  • 通过 Interop.UIAutomationClient 启动类型化 UI 自动化客户端

  • 读取丰富的桌面根和焦点元素元数据，包括边界、运行时 ID 和支持的模式

  • 从原生窗口句柄和屏幕坐标解析元素

  • 按名称、类名、自动化 ID、框架 ID、控件类型和进程 ID 搜索 UIA 树

  • 从原始、控件或内容树视图枚举直接子元素

  • 从原始、控件或内容树视图枚举后代

  • 使用原始、控件和内容遍历器导航 UIA 树

  • 支持检查、搜索和导航的构建缓存请求

  • 读取文本、选择和键模式状态

  • 等待焦点、自动化、属性更改和结构更改事件

  • 执行常见操作，如焦点、调用、设置值、切换、展开/折叠、选择、窗口状态更改、移动/调整大小、滚动和范围值更新

  • 通过 CLI 和 MCP 公开相同的扩展表面

  • 将 CLI 和 MCP 服务器打包在 VS Code 扩展中

仓库布局

  • src\UIAutomationMcp.ComInterop

  • src\UIAutomationMcp.Service

  • src\UIAutomationMcp.CLI

  • src\UIAutomationMcp.McpServer

  • src\UIAutomationMcp.Smoke

  • vscode-extension

  • mcpb —— Claude Desktop 捆绑包源和构建脚本

  • skills —— 用于 uia-cli 和 uia-mcp 的代理技能

构建和验证

dotnet build .\UIAutomationMcp.sln -c Release
.\src\UIAutomationMcp.Smoke\bin\Release\net9.0-windows\UIAutomationMcp.Smoke.exe
.\src\UIAutomationMcp.CLI\bin\Release\net9.0-windows\uiamcp.exe inspect --root --cache
.\src\UIAutomationMcp.CLI\bin\Release\net9.0-windows\uiamcp.exe find --focused --scope children --max-results 10 --cache
.\src\UIAutomationMcp.CLI\bin\Release\net9.0-windows\uiamcp.exe children --focused --view raw --max-results 10
.\src\UIAutomationMcp.CLI\bin\Release\net9.0-windows\uiamcp.exe descendants --focused --view raw --max-results 25
.\src\UIAutomationMcp.CLI\bin\Release\net9.0-windows\uiamcp.exe wait-event --event-kind focus --timeout-ms 500
.\src\UIAutomationMcp.CLI\bin\Release\net9.0-windows\uiamcp.exe action focus --focused

相关文档

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityStale
ResponsivenessNo recent community issues

Resources

Related MCP Connectors

Related MCP Servers

  • Little Mouse Pointer

    Fish121380
    A
    license
    Not graded
    quality
    B
    maintenance
    Enables AI agents to select Windows desktop UI elements, windows, or screen regions via mouse hover, then obtain context through UI Automation, screenshots, and local OCR for MCP-compatible clients.
    2
    MIT

  • windows2text

    ArronHC
    A
    license
    Not graded
    quality
    C
    maintenance
    Enables Windows UI Automation screen-to-text and computer-use via MCP, allowing agents to observe, decide, and act on Windows applications through accessibility data without OCR.
    1
    MIT

  • Windows-Use MCP Server

    kultur-512
    A
    license
    Not graded
    quality
    C
    maintenance
    Enables MCP clients to observe and control the Windows desktop through the UI Automation accessibility tree — no vision model required — by exposing tools for reading windows and interactive elements, clicking, typing, scrolling, shortcuts, PowerShell, screenshots, virtual desktops, web scraping, file operations and persistent memory, including background-window control without stealing focus. An optional agent mode adds a single natural-language task tool that runs the full upstream agent loop when an LLM provider is configured.
    1
    MIT

  • Deskhand-mcp

    irmia2026
    A
    license
    Not graded
    quality
    C
    maintenance
    Enables an MCP host to observe and control Windows desktop applications deterministically: reading UIA control trees, OCR text and pixel diffs, clicking, typing, scrolling and window management, plus verbatim text reading/appending via ValuePattern/TextPattern. Visual judgment is delegated back to the host model's own multimodal context through annotated screenshots, numbered element cards, grid maps and crosshair zoom-ins, so no separate vision model is needed.
    1
    AGPL 3.0