Azure MCP Server

Official Microsoft MCP server that gives AI coding agents natural-language access to 45+ Azure service areas.

Description

The Azure MCP Server bundles all of Microsoft's Azure MCP tools into a single server and has been generally available (GA) since version 2.0. It implements the Model Context Protocol to give AI agents a seamless connection to Azure services.

Per the official documentation, the server covers more than 45 Azure service areas, including virtual machines, storage, databases, and AI services, and lets you manage Azure resources through natural-language commands. It can be started locally via npx, dotnet tool, uvx, or Docker, or installed as an extension in VS Code, Visual Studio, IntelliJ IDEA, or Eclipse; integrations for GitHub Copilot CLI/SDK and cloud agents are also available.

Supported clients include VS Code and Visual Studio as well as IntelliJ IDEA, Eclipse, Cline, Windsurf, Claude Desktop, Claude Code, Google Antigravity, and Amazon Q Developer. Authentication uses the Azure CLI, the Azure Identity SDK, or environment variables (Azure Entra ID), and supports the public Azure cloud as well as the China and US Government clouds.

Starting via npx:

Code
npx -y @azure/mcp@latest server start

Why a bundled server instead of many separate tools

Before the Azure MCP Server reached GA status, teams often had to maintain separate, service-specific integrations for different Azure services. The bundled server solves this by combining every Azure MCP tool under a single process: a coding agent can access virtual machines, databases, and AI services alike with a single configuration, without setting up a separate connection for each service area. For teams managing several Azure services in parallel, this significantly reduces configuration overhead.

Typical use cases

In practice, the server suits tasks such as querying a subscription's resource inventory in natural language, creating or scaling resources directly from a conversation with the coding agent, debugging deployment issues using Azure diagnostic data, or compiling cost reports across multiple resource groups. For DevOps teams that already spend a lot of time in the Azure Portal UI or the Azure CLI, natural-language control can noticeably speed up recurring queries.

Multi-cloud support as a distinguishing feature

Besides the public Azure cloud, the server, per the documentation, explicitly supports the China cloud as well as the US Government cloud. This matters for organizations that, for regulatory reasons, cannot use the public Azure instance but still want the same AI-assisted workflow for their cloud environment. Authentication uses the same mechanisms across all three environments (Azure CLI, Azure Identity SDK, or environment variables).

Installation and Connection

You can install the Azure MCP Server as a package through NuGet (Azure.Mcp), npm (@azure/mcp) or PyPI (msmcp-azure). This is not required, because the described connections start it themselves. The documentation lists guides for Cline, Cursor, Google Antigravity, IntelliJ, Visual Studio, Visual Studio Code and Windsurf, as well as for GitHub Copilot CLI, SDK and cloud agent. It can also run as a Docker container and be addressed from Python and .NET through MCP client libraries.

Frequently asked questions

Does the server need its own Azure account with special permissions? No, it uses exactly the permissions of the signed-in account or service principal; the server itself grants no additional rights. Does the server work without Node.js? Yes, dotnet tool, uvx, or a Docker image are available alternatives. Is the server suitable for production environments? Per the provider, yes — it has been generally available (GA) since version 2.0, though security depends substantially on the account permissions granted.

Requirements

An Azure account with the permissions required for the services being accessed. Depending on the installation method, also Node.js (npx), .NET (dotnet tool), Python/uv (uvx), or Docker. Sign-in via the Azure CLI, the Azure Identity SDK, or environment variables is required.

Installation instructions

The Azure MCP Server can be set up in several ways — via an IDE extension, a package manager, or as a container. Which path fits best mainly depends on the client you use.

Step 1: Choosing an installation path

For VS Code, the Marketplace extension is the best fit since it bundles installation and sign-in into one step. For terminal/CLI agents such as Claude Code or Codex, starting via npx, dotnet tool, or uvx is more practical. Docker suits you if the server should run isolated or inside a CI/CD pipeline.

Setting up via VS Code

First install the "Azure MCP Server" extension from the VS Code Marketplace, then run "Azure: Sign In" via the Command Palette. The server is then usable for GitHub Copilot Chat and other MCP-capable tools in VS Code.

Step 2: Setting up via a package manager

Depending on your existing toolchain, pick one of the following:

bash
# Node.js
npx -y @azure/mcp@latest server start

# .NET
dotnet tool install Azure.Mcp

# Python (uv)
uvx --from msmcp-azure azmcp server start

# Docker
docker run -i --rm mcr.microsoft.com/azure-sdk/azure-mcp:latest

Sign in to Azure with az login before the first call.

bash
az login

Scoping permissions

The server's effective permissions exactly match the signed-in Azure account. For automated agents, a dedicated service principal with as narrow an RBAC role as possible is recommended, rather than a personal account with broad rights.

Step 3: Verifying the setup

After starting, test with a harmless question such as "What resource groups do I have in my subscription?" to confirm the agent can use Azure tools, and check the Azure Activity Log to see which actions are actually performed.

Step 4: Working with multiple Azure accounts

Anyone working across multiple Azure subscriptions or tenants should sign in separately for each account via az login --tenant <tenant-id> and confirm the active subscription before each session with az account show. An accidentally wrong active subscription can otherwise lead the agent to make changes in the wrong environment.

bash
az login --tenant <tenant-id>
az account show

Troubleshooting connection issues

If the client reports it can't connect to the server, this is often an expired Azure CLI session — running az login again usually resolves it. With the VS Code extension, restarting the editor after signing in often helps the extension pick up the current session.

bash
az login

Step 5: When Docker beats a local installation

For CI/CD pipelines or isolated test environments, the Docker route is often preferable, since it requires no local installation of Node.js, .NET, or Python and integrates easily into existing container workflows. Credentials must reach the container via environment variables or mounted configuration files, never baked into the image itself.

Source: learn.microsoft.com/en-us/azure/developer/azure-mcp-server, checked on 2026-09-05.

npx -y @azure/mcp@latest server start

Authentication

Azure CLI (az login), the Azure Identity SDK, or environment variables for Azure Entra ID. Supports Azure Public Cloud, Azure China Cloud, and Azure US Government Cloud.

Required access permissions

Effective permissions exactly match those of the signed-in Azure account or service principal — the server itself grants no additional rights. Depending on role assignment, read and write access to any of the 45+ supported Azure service areas is possible.

Transmitted or stored data

Requests and tool parameters go to the respective Azure services via the official Azure SDKs. Responses can include resource data from any Azure services used, and are passed to the MCP client and its language model. With local installation (npx, dotnet, Docker), the server runs on your own machine; access and audit logging occur as usual in Azure Monitor/Activity Log.

Security risks

Because the server can call any of the 45+ Azure service areas, an overly broad Azure account or a compromised host enables far-reaching actions in the subscription, up to changes on production resources. Grant only minimal, task-scoped RBAC roles, use dedicated service principals rather than personal accounts for automated agents, and monitor the Azure Activity Log for unusual activity.

License and costs

License
MIT
Cost
free

The Azure MCP Server itself is free and open source (MIT license). Only the Azure resources actually used through the server are billed, at regular Azure rates.

Alternatives

Not recorded yet.

At a glance

Provider
Microsoft
Status
Official server
Deployment
Local and remote
Current version
Not recorded yet.
GitHub stars
3,859
Last reviewed
07.09.2026

Repository and documentation

Categories

Supported clients