This MCP server package provides a central MCP Registry that coordinates multiple MCP servers and a Python-based HTTP client/server for interacting with MCP tools. The registry handles registration, health monitoring, and unified access to tools across connected MCP servers. Use the HTTP components to expose MCP tools over HTTP, enabling both synchronous and asynchronous interactions, and leverage the MCP Explorer UI for graphical interaction with registered servers and their tools. The FastMCPHttpServer lets you define tools that can be invoked remotely, while the FastMCPHttpClient can Discover servers, list available tools, and call tools via HTTP endpoints. This setup is useful for orchestrating multiple MCP-enabled services behind a single registry and providing a consistent HTTP API for tools and resources.

Prerequisites:

• Python 3.8+ installed on your system
• git to clone the repository
• Network access to install Python dependencies

1. Clone the repository git clone https://github.com/your-org/mcp-registry.git cd mcp-registry

2. Create and (optionally) activate a virtual environment python -m venv venv

   On Windows

   venv\Scripts\activate

   On macOS/Linux

   source venv/bin/activate

3. Install dependencies pip install -r requirements.txt

4. Run the registry server (as described in the README usage) python start_registry_server.py

5. Start an MCP server and register it with the registry (example usage from the README) python -c "<example or server start command>" # or run your MCP server script

6. Use the HTTP client to interact with the registry and registered MCP servers python -m pip install --upgrade fastmcp_http # if using the HTTP client utilities

Notes:

• The registry port defaults to 31337 but can be configured via the REGISTRY_PORT environment variable when launching the server.
• Ensure network accessibility between the registry and any MCP servers you register.

Tips:

• If you run into port conflicts, change the REGISTRY_PORT environment variable or the port in the registry startup script.
• Make sure all MCP servers you want to manage are reachable from the registry and properly registered on startup.
• The Python HTTP client/server components rely on the same MCP protocol; use the FastMCPHttpClient to list servers/tools and to call tools remotely.
• For deployment, consider securing the HTTP endpoints and using proper authentication if exposing the registry publicly.