mirzaasadijaz/enterprise-ai-agent repository overview

This image provides an AI shopping assistant for a luxury fashion brand using an agentic RAG system. It utilizes a LangGraph ReAct agent powered by Groq to intelligently route queries between internal document search and live web search.

Key features included in this image:

• Agentic Routing: Uses a LangGraph ReAct agent to dynamically choose between searching uploaded internal documents via ChromaDB or the live web via Tavily.
• Advanced RAG Pipeline: Supports ingesting multiple file formats (PDF, DOCX, CSV, TXT, MD) which are chunked, embedded locally using sentence-transformers, and stored in ChromaDB.
• Multi-User Memory: Implements persistent conversation history using SQLite, assigning unique session IDs to each user's LangGraph thread.
• Brand-Aware Persona: Configured with a specialized system prompt to maintain a polite, luxurious tone that prioritizes business and never recommends competitors.
• Included Interfaces: Ships with a FastAPI-powered backend serving HTML chat and upload pages, along with an optional Streamlit testing UI.