Supercharging Visual Studio Code with Real-Time AI Knowledge: How MCP and Microsoft Learn Transform Your Assistant
By Ivana Tilca · July 1, 2025 · 5 min read
By combining the Model Context Protocol with real-time documentation search, Microsoft has given AI assistants an operational memory—rooted in truth, context, and immediacy. In Visual Studio Code, this translates to less context switching, more accurate answers, and an assistant that doesn't just complete your code—but truly understands your workspace and your goals. MCP isn’t just a protocol. It’s the beginning of AI that acts on your behalf—with precision, depth, and practical value.
In today's AI-powered development workflows, context is everything. Microsoft’s Model Context Protocol (MCP) changes the game by letting AI assistants go beyond general answers—they can now interact with real tools, services, and documentation as you code.
One standout example is the Microsoft Learn Docs MCP Server, which brings official, real-time documentation into your AI's fingertips—directly inside Visual Studio Code. That means your assistant doesn’t just suggest code—it can query Microsoft Docs live, cite the source, and give you deeply relevant info for what you're building.
Let’s break down what MCP is, how it works inside VS Code, and why this opens up a whole new level of productivity.
https://www.youtube.com/watch?v=L4ONs-sRSD0
What is MCP(Model Context Protocol)?
MCP is a standardized system that enables AI models to discover, understand, and communicate with external tools, data sources, and applications. Imagine handing your AI a toolbox—along with instructions on which tool to use, how to use it, and when.
When an AI assistant receives a prompt—say, inside Visual Studio Code—MCP helps it determine:
🔧 What tools exist (e.g., file readers, database checkers, documentation search)
📘 How to use each tool (parameter structure, API style, input/output schema)
🚦 When to invoke them, so responses go beyond generic knowledge and take action behind the scenes
This allows the assistant to dynamically open files, fetch data, or call services as needed—then deliver smarter, more relevant responses.
How does it work?
MCP creates a shared language between the AI assistant and external tool providers—called MCP servers.
Here’s how the process feels in action:
👨💻 The user makes a request (e.g., “check recent API changes”)
🤖 The AI determines that an external tool could help
🛠️ The MCP server receives the request, processes it (like searching documentation), and returns structured results
To make this smooth, MCP defines how each tool should be described, how it should be invoked, and how its results are reported—whether the connection is via polling, streaming, or HTTP-based interaction.
MCP in Visual Studio Code
Visual Studio Code plays a starring role by acting as a gateway for this whole experience. It allows AI-powered assistants to communicate with MCP servers via:
📡 Walkie-talkie-style messaging (back-and-forth requests)
🔁 Live streamed updates (like a sports ticker)
🌐 Web-style message exchanges (REST-like behavior)
Beyond messaging, VS Code also informs the MCP server about the project workspace—letting tools operate with more context. It’s all designed to make the AI feel like an integrated teammate, not a disconnected chatbot.
Microsoft Learn Docs MCP Server
One of the most powerful use cases for MCP in VS Code is Microsoft’s official Learn Docs MCP Server. Here’s why it stands out:
-It enables AI assistants (like GitHub Copilot Chat) to search official Microsoft documentation across Learn, Azure, and M365 in real time.
-It returns structured responses with titles, summaries, and source URLs.
-It ensures reliability and trust—no hallucinated answers, just grounded facts.
Behind the scenes, it connects via a service endpoint: https://learn.microsoft.com/api/mcp (This is for machine-to-machine interaction—not something you can open in a browser.)
So whether you need to understand Azure Key Vault or find the latest deployment strategies, your AI taps right into the same documentation you would manually search—only faster.
How to add an MCP Server Globally in VS Code
Enable MCP support in VSCode.
To enable MCP support in VS Code, enable the chat.mcp.enabled setting.
1- Open VS Code Launch Visual Studio Code on your machine.
2- Go to Settings
-Press Ctrl + , (Windows/Linux) or Cmd + , (macOS)
-Or click the gear icon ⚙️ in the lower-left corner and choose Settings
3- Search for the Setting In the search bar at the top of the Settings panel, type:
chat.mcp.enabled
4- Enable the Setting You should see a checkbox labeled “Enable MCP support in chat” (or similar). ✅ Make sure it’s checked.
5 - Restart VS Code (if needed) Some features may require a quick restart to fully activate.
Let’s add Microsoft Learn Docs MCP Server to VS Code
1- Go to Settings
-Press Ctrl + , (Windows/Linux) or Cmd + , (macOS)
-Or click the gear icon ⚙️ in the lower-left corner and choose Settings
2 - In your settings.json file, insert this block (or merge it if you already have other configurations):
"mcp": { "servers": { "microsoftDocs": { "type": "http", "url": "https://learn.microsoft.com/api/mcp" } } }
3- Reload VS Code. Run Developer: Reload Window from the Command Palette or fully close and reopen VS Code.
4- Confirm the Server is Active. Open Command Palette (Ctrl + Shift + P)
Run:
MCP: List Servers: You should see microsoftDocs in the list 🎉
5- As you can see in the image, the server is stopped, go ahead and select it to start the server.
Using It Inside Chat
1- Activate Agent Mode. Open the Copilot Chat view (Ctrl + I or ⌃⌘I). From the dropdown menu above the input box, select Agent
2- Click on configuration tools and the MCP server will be included.
3 - 💬 When you ask a question (e.g. “What is Azure Key Vault?”), the AI runs that MCP tool
📎 You get an answer with verified Microsoft Learn content, extracted and summarized for you
And at the end you will have:
The integration is deep, smooth, and incredibly useful for developers who want insights without losing focus or flipping tabs.
Conclusion
By combining the Model Context Protocol with real-time documentation search, Microsoft has given AI assistants an operational memory—rooted in truth, context, and immediacy.
In Visual Studio Code, this translates to less context switching, more accurate answers, and an assistant that doesn't just complete your code—but truly understands your workspace and your goals.
MCP isn’t just a protocol. It’s the beginning of AI that acts on your behalf—with precision, depth, and practical value.