Multi-Tenant RAG System

A Retrieval-Augmented Generation (RAG) chat application designed for company knowledge management. This application allows employees to query company documents through an intelligent AI-powered chat interface, with department-specific access controls (similar to RBAC) and secure hashing-based authentication.

GitHub

Multi-Tenant RAG System

A Retrieval-Augmented Generation (RAG) chat application designed for company knowledge management. This application allows employees to query company documents through an intelligent AI-powered chat interface, with department-specific access controls (similar to RBAC) and secure hashing-based authentication.

Architecture

Features

Core Features

  • Secure Authentication: Employee login with department-based access control
  • Department-Specific Knowledge Access: Users can only access documents relevant to their department
  • AI-Powered Chat Interface: Intelligent responses using Google Gemini 2.5 Pro with context from company documents
  • Hybrid Search: Combines semantic and BM25 search for accurate document retrieval
  • Memory Support: Maintains conversation context using LangChain memory buffers
  • Modern UI: Beautiful, responsive interface with frosted glass design and gradient backgrounds
  • Real-time Chat: Instant responses with loading states and error handling

Technical Features

  • Vector Database Integration: Uses Qdrant for efficient document storage and retrieval
  • OpenAI Embeddings: Text embeddings using OpenAI's text-embedding-3-small model
  • Workflow Automation: n8n workflows for login validation and chat processing
  • PostgreSQL Integration: Secure user credential storage with password hashing
  • Evaluation System: Built-in response quality evaluation using Google Sheets integration

Read more on GitHub.