Agentic Product App

About

This project offers a complete voice and natural language interface for product catalog search and AI-driven chat assistance. It leverages Next.js as an MCP Host for LLM orchestration and Whisper transcription, alongside a FastAPI backend serving as a pure MCP Tool Server that provides structured product functions, semantic search, and Retrieval-Augmented Generation (RAG) capabilities. The architecture highlights the Model Context Protocol (MCP) by demonstrating how domain-specific tools remain decoupled from the host's AI and voice layers, creating a reusable and extensible backend for intelligent product interactions.

Key Features

• Model Context Protocol (MCP) Integration for decoupled components
• 0 GitHub stars
• Automatic Function Calling via Google Gemini 2.0 Flash
• AI Chat Assistant with RAG capabilities
• Voice and Natural Language Product Search
• In-Memory Vector Search for fast semantic product discovery

Use Cases

• Creating conversational interfaces for complex product data
• Building AI-powered e-commerce or product catalog assistants
• Demonstrating Model Context Protocol principles in action