This MCP server is a lightweight FastAPI-based implementation of the Model Context Protocol (MCP). It exposes two endpoints: a health check at GET / and a context processor at POST /context. The server accepts a JSON payload containing a prompt_id and optional parameters, and returns a generated context based on the provided template. It is designed to demonstrate basic MCP concepts such as template-driven prompts and parameter substitution. To run locally, start the server and then call the /context endpoint with a JSON payload specifying the prompt_id and any required parameters. The example shows a greeting prompt with a time parameter, illustrating how parameterized prompts can be supplied to customize responses.

Prerequisites:

• Python 3.9 or newer
• pip (comes with Python)
• Network access to install dependencies

Step-by-step installation:

1. Create and activate a virtual environment (recommended): """ python -m venv venv

On Windows:

venv\Scripts\activate

On macOS/Linux:

source venv/bin/activate """ 2) Install dependencies from requirements.txt: """ pip install -r requirements.txt """ 3) Run the server (as in the README): """ uvx src.main:app --reload """ 4) Open http://localhost:8000 to verify the health check and use the /context endpoint.

Notes and tips:

• The server runs by default on http://localhost:8000. If you need to change the port, configure the uvicorn/uvx command accordingly.
• The /context endpoint expects a POST with a JSON body containing at least a prompt_id, and optionally a parameters object for templated prompts, e.g., {"prompt_id": "greeting", "parameters": {"time": "12:00 PM"}}.
• Ensure requirements.txt is present and contains all necessary dependencies for FastAPI and your context logic.
• If you encounter CORS or network issues, ensure the server is reachable from the client and that any firewalls allow port 8000.
• For development, use the --reload flag as shown to auto-reload on code changes (works with uvx as wrapper).