LinkedIn MCP Server

# MCP FastAPI Client This is a FastAPI client for interacting with Model Context Protocol (MCP) servers. It provides a REST API interface to test and interact with MCP servers. ## Features - Server information retrieval - List available tools - Call tools with parameters - List available resources - Read resources ## Prerequisites - Python 3.7+ - pip ## Installation 1. Install dependencies: ```bash pip install -r requirements.txt ``` 2. Configure the environment variables in the `.env` file: ``` # The URL of the MCP server to connect to MCP_SERVER_URL=http://localhost:8888/mcp # The base URL for the API client (used by test_client.py) API_BASE=http://localhost:8001 ``` Replace these values with your actual URLs if they're different. ## Usage ### Managing Services The following scripts are provided to help manage the MCP server and FastAPI client: - **start.sh**: Starts both the Netlify MCP server and FastAPI client in the background ```bash ./start.sh ``` - **stop.sh**: Stops both services properly ```bash ./stop.sh ``` - **check_status.sh**: Checks the status of both services and displays logs ```bash ./check_status.sh ``` - **test_client.py**: Runs tests against the FastAPI client ```bash ./test_client.py ``` ### Starting the Server Manually If you prefer to start the FastAPI client manually: ```bash uvicorn main:app --reload --port 8001 ``` The API will be available at http://localhost:8001 ### API Documentation Interactive API documentation is available at http://localhost:8001/docs ### Troubleshooting If you encounter issues: 1. Check if the ports are already in use (8001 for FastAPI, 8888 for Netlify) 2. Verify both services are running with `./check_status.sh` 3. Check the log files in the `.processes` directory 4. Ensure your `.env` file has the correct URLs 5. Try stopping and restarting both services with `./stop.sh` followed by `./start.sh` ### API Endpoints #### Server Information ``` GET /server ``` Returns information about the MCP server. #### List Tools ``` GET /tools ``` Returns a list of available tools on the MCP server. #### Call Tool ``` POST /tools/call ``` Calls a specific tool with the provided arguments. Example request body: ```json { "name": "run-analysis-report", "args": { "days": 5 } } ``` #### List Resources ``` GET /resources ``` Returns a list of available resources on the MCP server. #### Read Resource ``` POST /resources/read ``` Reads a specific resource. Example request body: ```json { "uri": "docs://interpreting-reports" } ``` ## Using with Different MCP Servers To use the client with a different MCP server, update the `MCP_SERVER_URL` in the `.env` file or set the environment variable before starting the server: ```bash export MCP_SERVER_URL=https://your-mcp-server-url/mcp uvicorn main:app --reload ```