mcp -with-streamable-http-example

This MCP server provides a simple streamable HTTP transport interface with a small set of example tools and prompts. It runs a Python-based MCP server that serves endpoints to interact with tools like hello_world, add_numbers, random_number, return_json_example, and calculate_bmi. You can customize the server port and enable verbose debug output via environment variables. The included tools demonstrate basic arithmetic, string output, random number generation, and a BMI calculation workflow, all accessible through the MCP protocol over HTTP.

Prerequisites:

• Python 3.x installed
• Access to run scripts from the repository
• Optional: uv for running the server via UV if you prefer the uv runtime

Installation steps:

1. Ensure you have Python installed and available in PATH.
2. Download or clone the repository containing simple_streamable_http_mcp_server.py.
3. Install any required Python dependencies if provided (usually via pip install -r requirements.txt if present).
4. Run the server using the instructions below.

Usage examples:

• Default port (8000): python simple_streamable_http_mcp_server.py
• Custom port (e.g., 9000): MCP_SERVER_PORT=9000 python simple_streamable_http_mcp_server.py
• Enable debug logging: MCP_DEBUG=1 python simple_streamable_http_mcp_server.py
• Using uv if preferred: uv run mcp-server

Notes and tips:

• The server defaults to port 8000 unless overridden with MCP_SERVER_PORT.
• Debug logging (MCP_DEBUG) is helpful for tracing tool invocations and inputs/outputs.
• The tools provided are examples: hello_world, add_numbers, random_number, return_json_example, and calculate_bmi. They illustrate typical MCP tool interfaces (name, parameters, and description).
• If you modify the script, ensure the port configuration and environment variables align with your deployment environment.
• Since this is a Python-based MCP server, npm-related configuration is not applicable.