Microsoft runs a free MCP server over its Learn documentation. An agent can search Azure, .NET and Microsoft 365 docs, read whole pages and pull official code samples, instead of answering from training data that predates the current SDK.
https://learn.microsoft.com/api/mcp
Gemini 3.8 Flash
Chat with 60+ AI models on the same workflow — switch to a different model mid-conversation and re-run the same prompt, or use Compare mode to put several side-by-side and balance quality vs. cost.
No token required
How models use it and what it is built for.
The server exposes three tools. microsoft_docs_search returns up to ten content chunks of at most 500 tokens each from Microsoft Learn and other official sources, each with its title and URL. microsoft_docs_fetch reads a full documentation page as markdown, for when a search excerpt cuts off half a procedure. microsoft_code_sample_search finds code snippets from the docs, optionally filtered by language. The value is currency: Azure services and SDKs change faster than any model’s training cutoff, so an agent that checks the docs before writing a Bicep file or a Graph call gets current parameter names rather than plausible old ones. Search first, then fetch the one page that matters, keeps the context small.
Typical tools an AI model can call. Exact names vary by version.
Taken from the official Microsoft Learn documentation — see Microsoft Learn MCP Server documentation for the full reference.
Client configuration
Any MCP client
No credential block needed — just the URL. It is an MCP endpoint, so opening it in a browser will not show a page.
{
"mcpServers": {
"microsoft-learn": {
"type": "http",
"url": "https://learn.microsoft.com/api/mcp"
}
}
}Copy any of these into MCP Agent Studio after connecting.
What is the current way to authenticate an Azure Function to Key Vault with a managed identity? Show the official sample.
Find the Microsoft Graph call to list a user’s calendar events in C#, from the official docs.
Fetch the Azure Container Apps scaling rules page and summarise the KEDA scalers it supports.
Which .NET version dropped support for this API, according to the docs?
This is not a single-model product: you get the same MCP connection with 60+ models (Claude, GPT, Gemini, DeepSeek, open-weight, and more), you can switch mid-conversation, and you can open Compare mode to run the same prompt against multiple models at once. The card above is a suggested starting point for this server — not the only choice.
Default pick for Microsoft Learn
Gemini 3.8 Flash
The pattern is search, read one page, answer. A fast model with a large window handles a full fetched page comfortably and keeps each turn cheap.
Listing tools proves the server is reachable, not that a model can work with it. Evals go further: they read every tool on the server, write a test suite from its real schemas, and run it — code decides pass/fail on the responses (schema conformance, error codes, pagination, result caps) while a scoring model grades plain-English tasks driven through the tools.
Get a pass/fail report per tool with the evidence behind each verdict — and replay the same suite after every schema change. Destructive tools are excluded from the run.
Run evalsOpen MCP Agent Studio with the connection pre-filled. Add your token, pick any of 60+ models, and start chatting — no install required.
Open Agent StudioCommon questions about connecting, scoping and using it safely.
Yes. It needs no Microsoft account or API key, and the hosted endpoint at https://learn.microsoft.com/api/mcp answers an initialize handshake unauthenticated.
Three: microsoft_docs_search for excerpts from official docs, microsoft_docs_fetch to read a full page as markdown, and microsoft_code_sample_search for official code snippets.
No. It searches Microsoft Learn broadly, including Azure, .NET, Microsoft 365 and other Microsoft technologies.
No. It only reads public documentation. To act on your subscription, use the Azure MCP server, and keep this one alongside it so the agent checks the docs before it changes anything.
It is an MCP endpoint, not a web page. It expects a JSON-RPC request from an MCP client, so a plain GET from a browser is rejected.
Azure
Manage Azure resources across 50+ services in natural language.
DeepWiki
Ask questions about any public GitHub repository and get answers grounded in its generated wiki.
Context7
Give your AI agent up-to-date, version-specific library docs.
AWS
Query AWS docs, pricing and infrastructure through an AI agent.