The project demonstrates how to create a custom MCP server and client from the ground up without relying on third-party libraries. It also illustrates how to integrate this client with an AI agent.
This project is based on a couple of YouTube videos by Jack Herrington:
This MCP server implements a coffee shop interface that allows AI agents to:
- Browse available coffee drinks (Latte, Mocha, Flat White)
- Get detailed information about each drink including prices and descriptions
The coffee shop MCP provides a practical example of building custom MCP tools that integrate with AI agents for real-world applications.
- Node.js - JavaScript runtime.
- TypeScript - A superset of JavaScript that adds static types.
- Model Context Protocol - The protocol for building AI applications.
For detailed instructions on interacting with each component:
- MCP Server README - Learn how to run and configure the MCP server, use the inspector interface, and connect external clients.
- MCP Client README - Discover how to set up AI agent connections, run the client, and use the interactive terminal interface.
