RAG

Corrective RAG Agent

A sophisticated LangGraph system that combines retrieval, relevance grading, and web-search to provide highly accurate responses.

Demo 2 of 15 Client Demo Ready

Quick Start

$ streamlit run corrective_rag.py

Key Features

Smart Document Retrieval: Uses Qdrant vector store for high-precision matching
Document Relevance Grading: Claude 4.5 Sonnet filters out irrelevant noise
Query Transformation: Automatically optimizes user queries for better search context
Web Search Fallback: Integrates Tavily API when internal documentation is insufficient

"Provide a research paper and ask specific questions; the agent will verify internal facts or find updates online."

Previous Demo

AI System Architect Advisor (R1)

All Demos

Back to Showcase

Next Demo

Multi-MCP Agent Router