The Azure MCP Server is a newly introduced, open-source solution that enables AI agents to interact seamlessly with Azure cloud resources using the Model Context Protocol (MCP). This protocol standardizes communication between agents (clients) and external resources (servers), allowing developers to build AI-driven workflows that can access, query, and manage Azure services with natural language commands.

What is the Azure MCP Server?

• MCP (Model Context Protocol): An open protocol designed to connect AI agents with external tools and data sources in a unified, scalable way. MCP uses a client-server architecture based on JSON-RPC 2.0, supporting transports like standard IO and HTTP with Server-Sent Events (SSE).

• Azure MCP Server: Implements the MCP standard to expose Azure services—such as file storage, databases, log analytics, and CLI operations—to AI agents. This enables agents to perform context-aware operations on Azure resources, such as querying databases, managing storage, or executing CLI commands.

Supported Azure Services and Tools (Public Preview)

The Azure MCP Server currently supports a broad set of Azure services and developer tools, including:

• Azure Cosmos DB: List accounts, query databases, manage containers/items, execute SQL queries.

• Azure Storage: List accounts, manage blob containers/blobs, query storage tables, retrieve metadata.

• Azure Monitor (Log Analytics): List workspaces, query logs with KQL, manage tables, configure monitoring.

• Azure App Configuration: List stores, manage key-value pairs, handle labeled configurations, lock/unlock settings.

• Azure Resource Groups: List and manage resource groups.

• Azure CLI: Execute any Azure CLI command, with JSON output support.

• Azure Developer CLI (azd): Run azd commands for template management, provisioning, and deployment.

Example Use Cases

Below are practical examples of what you can do with the Azure MCP Server:

• Explore Azure Resources:

  • “List my Azure storage accounts”

  • “Show me all my Cosmos DB databases”

  • “List my resource groups”

  • “Show me the tables in my Storage account”

• Query and Analyze:

  • “Query my Log Analytics workspace”

• Manage Configuration:

  • “List my App Configuration stores”

  • “Show my key-value pairs in App Config”

• Advanced Operations:

  • “List my Azure CDN endpoints”

  • “Help me build an Azure application using Node.js”

How to Use the Azure MCP Server

For GitHub Copilot Agent Mode:

• Install the Azure MCP Server via the GitHub repository (one-click install available).

• In VS Code, switch GitHub Copilot to Agent Mode and issue natural language prompts (e.g., “List my Cosmos DB accounts”).

• For deeper integration, combine with the GitHub Copilot for Azure extension for enhanced documentation and resource graph queries.

For Custom MCP Clients:

• nstall and start the server using Node.js:

  npx -y @azure/mcp@latest server start

  For SSE transport (networked/remote usage):

  npx -y @azure/mcp@latest server start --transport sse

  Configure your MCP client to connect via the appropriate transport and port (default: http://localhost:5008/sse)].

For Azure Functions MCP Server:

• Use Azure Functions with specialized MCP triggers and bindings to create scalable, event-driven tools accessible by AI agents.

Integration Example

Integrating with Claude Desktop (or other MCP clients):

• Start the Azure MCP Server.

• Update the client configuration to include the MCP server endpoint and any required environment variables (e.g., connection strings, agent IDs).

• Use provided methods, such as list_agents(), connect_agent(), or query_default_agent(), to interact with Azure resources through the MCP interface.

Summary Table: Example Prompts and Corresponding Azure Services

Getting Started

• Install Node.js (required for the Azure MCP Server).

• Clone the Azure MCP Server repository or use the provided npm command.

• Connect your AI agent (e.g., GitHub Copilot, custom MCP client) to the server using standard IO or SSE.

• Issue natural language commands to perform operations on your Azure resources.

The Azure MCP Server is currently in Public Preview, with ongoing development and more features expected soon. It represents a significant step toward agentic, AI-driven cloud development, making Azure resources accessible and manageable through conversational interfaces and standardized protocols