AI Todo

소개

This project showcases an API-first approach to the Model Context Protocol (MCP), featuring a FastAPI backend that simultaneously exposes both a traditional REST API and a dedicated MCP endpoint. It provides a robust todo service where the backend acts as the single source of truth, offering its capabilities through both REST endpoints and MCP tools. A Next.js frontend, designed to be thin, consumes the REST API for its user interface while also enabling automated workflows via MCP. This architecture allows assistants within MCP-compatible clients to directly invoke the same server-side tools utilized by the UI, streamlining development and ensuring consistent logic across different interaction paradigms.

주요 기능

• API-first architecture combining FastAPI backend and Next.js frontend
• MCP endpoint enabling AI assistants to invoke server-side todo management tools
• Next.js client with multi-tab UI for CRUD operations and an AI chat interface
• Dual exposure of backend logic via REST API and MCP tools
• Chat feature displaying per-response tool-call details for transparency
• 0 GitHub stars

사용 사례

• Orchestrating backend services through both traditional UIs and AI assistants
• Demonstrating an API-first strategy for Model Context Protocol implementation
• Building AI-powered applications with integrated tool-calling capabilities