VS Code tools for Windows development

Two VS Code tools round out the AI-assisted Windows development workflow: the WinApp extension brings the Windows App Development CLI into the editor, and the Microsoft Learn MCP Server gives your AI agent live access to current Windows documentation.

WinApp VS Code extension

The WinApp extension brings the Windows App Development CLI into VS Code — initialize, run, debug, package, and sign Windows apps without leaving the editor.

Install

code --install-extension microsoft-winappcli.winapp

Or search WinApp in the Extensions panel (Ctrl+Shift+X). Requires the WinApp CLI to be installed first.

Command Palette commands

All commands are available via Ctrl+Shift+P → WinApp:

Workflow

1. dotnet new winui-navview -n MyApp — scaffold project
2. cd MyApp && dotnet run — build and verify it runs
3. code . — open in VS Code
4. Ctrl+Shift+P → WinApp: Run Application — run with package identity
5. Edit XAML and C# files with AI assistance
6. Ctrl+Shift+P → WinApp: Create MSIX Package — package for distribution
7. winapp store publish ./*.msix --appId <your-app-id> — publish to the Store

Microsoft Learn MCP Server

AI models are trained on a snapshot of the web. For Windows development, that means your agent may have learned from WPF and UWP samples written years before WinUI 3 existed — and it can't tell the difference. The Microsoft Learn MCP Server fixes this by giving your agent a tool it can call to retrieve current, authoritative documentation at the moment it needs it.

What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI agents call external tools and data sources during a conversation. Instead of relying entirely on training data, an MCP-connected agent can search and read live content — including Microsoft Learn — before generating a response.

Add the Microsoft Learn MCP Server

The server is hosted by Microsoft and requires no installation or sign-in.

VS Code (GitHub Copilot)

Add the following to .vscode/mcp.json in your project:

{
  "servers": {
    "microsoft-learn": {
      "type": "http",
      "url": "https://learn.microsoft.com/api/mcp"
    }
  }
}

VS Code will prompt you to enable the server the first time you open a Copilot chat session.

Claude Code

Add the server to your Claude Code configuration (~/.claude/mcp_servers.json):

{
  "microsoft-learn": {
    "type": "http",
    "url": "https://learn.microsoft.com/api/mcp"
  }
}

Other MCP clients

Any client that supports the MCP HTTP transport can connect using:

https://learn.microsoft.com/api/mcp

No API key or authentication required.

What the server can do

Once connected, your agent can search and retrieve pages from Microsoft Learn. For Windows development, this means it can look up:

• Current WinUI 3 control APIs and usage patterns
• Windows App SDK release notes and migration guides
• winapp CLI command reference
• Store submission requirements and certification criteria

Example

Without the MCP server, asking Copilot to add a file picker may produce code using the deprecated UWP FileOpenPicker pattern:

// ❌ UWP pattern — may be generated without MCP context
var picker = new FileOpenPicker();
picker.SuggestedStartLocation = PickerLocationId.DocumentsLibrary;

With the MCP server connected, the agent retrieves the current WinUI 3 guidance and generates the correct pattern:

// ✅ WinUI 3 pattern — retrieved from current docs
var picker = new FileOpenPicker();
var hwnd = WinRT.Interop.WindowNative.GetWindowHandle(this);
WinRT.Interop.InitializeWithWindow.Initialize(picker, hwnd);
picker.SuggestedStartLocation = PickerLocationId.DocumentsLibrary;
var file = await picker.PickSingleFileAsync();

Related content

• Quickstart: Build and publish a Windows app with AI
• WinUI agent plugin
• Model Context Protocol specification
• GitHub Copilot MCP documentation