Test MCP Server Using Streamable HTTP

This sample project demonstrates a minimal FastMCP server using the Streamable HTTP transport, which has become the recommended default for remote MCP servers. It exposes a single greet tool that returns a friendly greeting.

Files

• server.py: Defines and runs the FastMCP server. At runtime, it calls mcp.run(transport="http") to launch the server on 0.0.0.0:8000/mcp. Adjust host, port or path as needed.
• requirements.txt: Contains fastmcp as the only dependency.

Running Locally

Ensure you have Python 3.10+ installed and a virtual environment set up. Then install dependencies and run the server:

pip install -r requirements.txt
python server.py

The server will start at http://localhost:8000/mcp. You can inspect it using an MCP client or inspector tool.

Deploying to Render

1. Create a new repository in GitHub and add these files. Link it to a new Web Service on Render.
2. Environment variables: None are strictly required for this simple example, but you can add READ_ONLY=true or MCP_PORT as needed.
3. Build and start commands (Streamable HTTP):
   • Build command: pip install -r requirements.txt
   • Start command: python server.py
4. Once deployed, your MCP server endpoint will be https://<render-app-url>/mcp. Use this URL when registering the server as a connector in ChatGPT when that feature becomes available.

Why Streamable HTTP?

The Model Context Protocol has been transitioning away from Server-Sent Events (SSE) due to limitations in scalability, reliability and bidirectional communication. The Streamable HTTP transport uses a single endpoint for both sending and receiving messages, dynamically upgrading to streaming when needed, which simplifies connection management and enhances robustness.