Azure Model Context Protocol (MCP) Hub

What is Azure Model Context Protocol (MCP) Hub?

Azure Model Context Protocol (MCP) Hub is a collection of resources and samples for building, running, or reusing MCP servers on Azure to power AI agents. It is designed for developers using C#, Python, Java, or JavaScript who want to connect AI agents to real APIs via the Model Context Protocol.

How to use Azure Model Context Protocol (MCP) Hub?

Choose a language to host your own MCP server using Azure Functions (samples provided for C#, Python, TypeScript). Alternatively, build agents using official MCP SDKs or AI framework integrations (OpenAI Agents SDK, Semantic Kernel, LangChain.js, Spring AI, Azure AI Agents). You can also use community-maintained plug-and-play MCP servers for data and dev/infra APIs. Each path links to detailed repository or documentation.

Key features of Azure Model Context Protocol (MCP) Hub

• Run your own MCP server with Azure Functions in C#, Python, or TypeScript
• Build agents using official MCP SDKs for C#, Python, TypeScript, and Java
• Integrate with AI frameworks: OpenAI Agents SDK, Semantic Kernel, LangChain.js, Spring AI, Azure AI Agents
• Access plug-and-play MCP servers for data APIs (Redis, PostgreSQL, MySQL, MongoDB, ADX)
• Access plug-and-play MCP servers for dev/infra APIs (Azure CLI, Kubernetes, GitHub, Azure DevOps)
• Community-maintained servers are not reviewed, supported, or endorsed by Microsoft

Use cases of Azure Model Context Protocol (MCP) Hub

• Host a custom MCP server on Azure Functions to expose proprietary APIs to AI agents
• Build an AI agent with Semantic Kernel that queries an Azure PostgreSQL database via MCP
• Create a natural-language interface for Azure CLI using the community server
• Add Bing search and Azure AI Search capabilities to an Azure AI Agent via MCP
• Control Azure Kubernetes Service (AKS) clusters through an AI assistant

FAQ from Azure Model Context Protocol (MCP) Hub

What is the Model Context Protocol (MCP)?

MCP is an open protocol that standardizes how applications provide context and tools to LLMs. It enables AI agents to call real APIs. The full specification is available at modelcontextprotocol.io.

Do I need to host a server to use this hub?

You can either host your own MCP server using Azure Functions, or reuse community-provided plug-and-play servers linked in the hub. No server is required to start building agents with the SDKs.

What languages and runtimes are supported?

Official MCP SDKs are available for C#, Python, TypeScript, and Java. The hub provides Azure Functions samples for C#, Python, and TypeScript. AI framework integrations cover Python, C#, JavaScript, and Java.

Are the plug-and-play servers supported by Microsoft?

No. The listed plug-and-play servers are open source and maintained by the community. Microsoft does not review, support, or endorse them. Use at your own discretion.

What transports or authentication are used?

The hub does not specify transport or authentication details. Each MCP server (custom or community) may use different transports (e.g., SSE) and auth methods. Refer to the respective SDKs and server documentation for specifics.