ai-foundry-ai-agent-mcp-server-example

# AI Foundry Agent Service MCP Server This project provides a simple MCP server and client for interacting with Azure AI Agents using FastMCP. It demonstrates how to expose tools and agent invocation endpoints over SSE (Server-Sent Events). ## Features - **MCP Server**: Exposes tools for addition and invoking Azure AI Agents. - **MCP Client**: Connects to the server, lists available tools, and calls them. - **Azure AI Integration**: Uses Azure AI Agents and Azure Identity for secure access. - **Environment Configuration**: Loads credentials and configuration from environment variables. ## Requirements - Python 3.8+ - `fastmcp` - `python-dotenv` - `azure-ai-agents` - `azure-identity` Install dependencies: ```sh pip install fastmcp python-dotenv azure-ai-agents azure-identity ``` ## Usage ### 1. Configure Environment Create a `.env` file with the following variables: ``` PROJECT_CONNECTION_STRING=your-azure-ai-endpoint MODEL_DEPLOYMENT_NAME=your-model-name AGENT_ID=your-agent-id ``` ### 2. Start the MCP Server ```sh python simple-agent-mcp-server.py ``` This will start the server on `http://127.0.0.1:8000/sse`. ### 3. Run the MCP Client ```sh python simple-client-mcp.py ``` The client will connect to the server, list available tools, and demonstrate calling the `add` and `invoke_agent` tools. ## File Overview - `simple-agent-mcp-server.py`: Main MCP server exposing tools and agent invocation. - `simple-client-mcp.py`: Example client for interacting with the server. ## Example Output ``` Available tools: ['add', 'invoke_agent'] Result: 8 Agent response: Paris ``` ## License See [LICENSE.md](LICENSE.md) for license information.