Demo

소개

This project offers a practical demonstration of an MCP (Model Context Protocol) server built with the `http-mcp` Python package. It showcases how to expose custom tools and prompts to any MCP-capable client, running efficiently over both HTTP (via Starlette/Uvicorn) and standard I/O (stdio). The demo includes illustrative examples of various tools, such as retrieving weather, getting current time, accessing request context, and tracking called tools, alongside a templated prompt for generating advice. It serves as an excellent reference for developers looking to integrate AI agents with external capabilities using the MCP standard.

주요 기능

• Offers an example prompt demonstrating dynamic content generation based on user input.
• Exposes a variety of example tools, including data retrieval and access to incoming request context.
• Supports both HTTP (Starlette/Uvicorn) and stdio communication modes for MCP server deployment.
• Leverages Pydantic v2 for robust input and output validation of all exposed tool functions.
• Includes a simple context mechanism to track session-specific tool calls.
• 0 GitHub stars

사용 사례

• Testing and validating MCP client configurations and interactions with a live server.
• Demonstrating the capabilities and implementation of the Model Context Protocol (MCP).
• Serving as a foundational example for building custom MCP servers in Python.