What is RAG Chat?
RAG Chat is a modern chat application that uses Retrieval-Augmented Generation (RAG) to provide AI responses grounded in your uploaded documents.
Proof of Concept — This project was built as a proof of concept to demonstrate RAG capabilities using Convex and AI agents. It is not intended for production use without further development and testing.
How It Works
- Upload Documents: Add PDF, Markdown, or text files to your knowledge base
- Automatic Processing: Documents are parsed, chunked, and indexed with vector embeddings
- Ask Questions: Chat with the AI about your documents
- Contextual Answers: The AI retrieves relevant chunks and generates informed responses
Tech Stack
| Package | Description |
|---|---|
| Convex | Backend platform with real-time sync |
| @convex-dev/agent | AI agent framework for Convex |
| @convex-dev/rag | RAG for Convex |
| TanStack Router | Type-safe routing for React |
| React | UI library |
| Vite | Build tool and dev server |
| Tailwind CSS | Utility-first CSS framework |
| shadcn/ui | Re-usable UI components |
| OpenRouter | AI model provider |
Features
- Document Management: Upload, view, and delete documents
- Real-time Chat: Instant message updates with Convex subscriptions
- Source Citations: See which documents informed each response
- Model Selection: Choose from various AI models via OpenRouter
- Modern UI: Clean interface built with shadcn/ui components