RetrievAI¶
RAG-based document retrieval and chat application powered by LangChain.
Overview¶
RetrievAI enables you to upload documents and chat with them using AI-powered retrieval augmented generation (RAG). The system processes your documents, creates vector embeddings, and provides accurate, context-aware responses.
Iframe not loading? Watch the video here
Features¶
- Document Upload — Support for PDF, DOCX, TXT, and Markdown files
- Intelligent Chunking — Documents are split into semantic chunks for optimal retrieval
- Vector Search — ChromaDB-powered similarity search for relevant context
- Chat Interface — Conversational AI with source citations
Quick Start¶
# Clone and configure
git clone https://github.com/your-username/retrievAI.git
cd retrievAI
cp .env.example .env
# Start services
docker compose up -d
Access the application:
- Frontend: http://localhost:3000
- API: http://localhost:8000
- API Docs: http://localhost:8000/api/docs
Tech Stack¶
| Component | Technology |
|---|---|
| Backend | FastAPI, LangChain, Python 3.11+ |
| Frontend | React 18, TypeScript, Vite |
| Database | PostgreSQL 16 |
| Vector Store | ChromaDB |
| Cache | Redis |
| Infrastructure | Docker Compose, Nginx |