MCP---Agent-Starter-Kit

The MCP---Agent-Starter-Kit documentation server provides an accessible web interface to explore and test the multi-agent workflow capabilities built into the Starter Kit. It hosts the internal MCP documentation, examples of multi-agent orchestration using FastAPI, and references to how agents can communicate with each other and with OpenAI services. Once running, you can browse documented APIs, run sample endpoints, and inspect how agents are coordinated within the kit's examples. This server is especially useful for developers who want to understand the architecture, test agent interactions, and validate integration points before building their own bot flows.

Prerequisites:

• Python 3.7 or later
• pip (comes with Python)
• Optional: virtual environment tool (venv)

1. Clone the repository (or the portion containing the MCP Docs server): git clone https://github.com/fub05/MCP---Agent-Starter-Kit.git cd MCP---Agent-Starter-Kit

2. Create and activate a virtual environment (recommended): python -m venv venv source venv/bin/activate # macOS/Linux venv\Scripts\activate # Windows

3. Install dependencies (adjust path if the docs server lives in a subdirectory): pip install -r mcp_docs_server/requirements.txt

4. Set environment variables (optional but recommended):

   • OPENAI_API_KEY=your-openai-api-key
   • UVICORN_PORT=8000

5. Run the documentation server: uvicorn mcp_docs_server.app:app --reload --port $UVICORN_PORT

6. Open a browser and navigate to http://localhost:8000 to access the MCP docs server.

Tips:

• Ensure you have a valid OpenAI API key if you plan to use OpenAI-powered examples.
• If the server does not start, check that port 8000 is free or adjust UVICORN_PORT accordingly.
• The documentation server is primarily for exploring the Starter Kit's MCP concepts and agent orchestration examples; replace paths with the actual project structure if different.
• For production deployments, consider running behind a reverse proxy and using proper SSL/TLS configuration.
• If you encounter dependency conflicts, use a clean virtual environment and install dependencies from a dedicated requirements file within the mcp_docs_server directory.