UIAutomationMcp

io.github.trsdnPublished by GitHub account trsdnv1.0.2Updated Oct 25, 2026

Windows UI Automation MCP server for desktop inspection and element lookup. Windows only.

VerifiedSTDIODesktop onlyBrowser AutomationDeveloper Tools

Overview

AI-generated overview

A Windows-only MCP server that lets an assistant inspect and control desktop applications through Windows UI Automation.

What you can do
Inspect the desktop root and focused element, including bounds, runtime IDs and supported patterns, and resolve elements from window handles or screen coordinates. Search UI Automation trees by name, class name, automation id, framework id, control type or process id, and enumerate children, descendants and walkers across raw, control and content views. Read text, selection, caret position, annotation runs and tabular cell matrices, and wait for focus, property-changed or structure-changed events. Perform actions such as focus, invoke, set-value, toggle, expand/collapse, selection, window state changes, move/resize, scroll, range-value updates, view switching and MSAA default actions.
When to use it
Useful when an assistant needs to observe or drive Windows desktop applications that expose UI Automation, including Win32, MFC and installer windows through the legacy accessibility bridge, or to locate items in long virtualized lists without scrolling.
Requirements
Windows with an interactive desktop session; UI Automation cannot run headless. Installed as a local stdio process, for example via the npm package, the NuGet .NET tool, the VS Code extension, the Claude Desktop bundle or a standalone ZIP. No authentication, environment variables or headers are declared.
Before you install
The server can act on desktop applications: focus, invoke, set values, toggle, resize, scroll and change window state, which may alter or submit data in whatever app is targeted. It reads screen content and element metadata from the desktop, so avoid running it where sensitive windows are open.

Installation

In SourceWeft

  1. Open UIAutomationMcp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

UIAutomationMcp

[VS Code Marketplace] [License: MIT] [.NET] [Platform]

Windows UI Automation toolkit with three aligned entry points:

  • CLI (uiamcp)
  • MCP server (UIAutomationMcp.McpServer)
  • VS Code extension (uia-mcp)

Install

VS Code extension (recommended) — install trsdn.uia-mcp from the Marketplace (code --install-extension trsdn.uia-mcp), or install the UIAutomationMcp-<version>.vsix from the latest release. Self-contained; bundles the MCP server, the CLI, and both agent skills.

Claude Desktop — download uia-mcp-<version>.mcpb from the latest release and double-click it.

npm — uia-mcp ships the MCP server with the .NET runtime bundled, so it needs no prerequisites:

json
{  "mcpServers": {    "uia-mcp": { "command": "npx", "args": ["-y", "uia-mcp"] }  }}

NuGet .NET tools

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

Standalone ZIP — UIAutomationMcp-MCP-Server-<version>-windows.zip contains the self-contained MCP server and CLI; no .NET runtime required.

Requires Windows and an interactive desktop session. UI Automation cannot run headless.

Lineage

This repository is part of the broader MCP automation lineage and intentionally keeps attribution to earlier work by Stefan Brönner.

It also keeps the historical reference to Excel MCP as a related sibling project in that lineage, even though this repository itself is focused on Windows UI Automation.

What it does

  • boots a typed UI Automation client through Interop.UIAutomationClient
  • reads rich desktop-root and focused-element metadata, including bounds, runtime IDs, and supported patterns
  • resolves elements from native window handles and screen coordinates
  • searches UIA trees by name, class name, automation id, framework id, control type, and process id
  • enumerates immediate children from raw, control, or content tree views
  • enumerates descendants from raw, control, or content tree views
  • navigates the UIA tree with raw, control, and content walkers
  • supports build-cache requests for inspection, search, and navigation
  • reads text, selection, and key pattern states, including caret position, annotation runs such as spelling errors, and the containing document of an inline element
  • waits for focus, automation, property-changed, and structure-changed events
  • performs common actions such as focus, invoke, set-value, toggle, expand/collapse, selection, window state changes, move/resize, scroll, range-value updates, view switching, docking, realizing virtualized items, and MSAA default actions
  • finds items that a virtualizing list has not materialized yet by asking the container directly, so a search for the 250th row of a 300-row list works without scrolling
  • reports drag and drop state (isGrabbed, drop effects, grabbed items) so an in-progress drag can be observed, and resolves automation events to readable names such as Drag_DragStart
  • reads MSAA data through the LegacyIAccessible bridge, so Win32, MFC, and installer windows that expose no native UI Automation provider still report usable names, roles, states, and a runnable default action
  • reads tabular controls as a cell matrix with row and column headers
  • exposes the same expanded surface through CLI and MCP
  • packages the CLI and MCP server inside the VS Code extension

Repository layout

  • src\UIAutomationMcp.ComInterop
  • src\UIAutomationMcp.Service
  • src\UIAutomationMcp.CLI
  • src\UIAutomationMcp.McpServer
  • src\UIAutomationMcp.Smoke
  • vscode-extension
  • mcpb — Claude Desktop bundle sources and build script
  • skills — agent skills for uia-cli and uia-mcp

Build and validate

powershell
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

Related docs

Source: README.md at commit ec0ffb1

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.0.2LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts