Knowledge Repo - RAG Chatbot

I built Knowledge Repo RAG Chatbot to learn how Retrieval-Augmented Generation (RAG) applications work end-to-end. It’s a chatbot that answers questions using resources from my own knowledge repository.

This is a Node & React project has two flavors:

• Local stack: Qdrant (vector DB), Nomic embeddings (via vLLM), and a LLaMA model with GPT4All.
• Cloud stack: OpenAI embeddings, Qdrant Cloud, and GPT completions via OpenAI API.

Key features include:

• Data ingestion pipeline
• Real-time chat with token streaming
• Chat memory, history, and automatic title generation
• Automated build and deployment

Planned improvements:

• Pipeline to auto-update the vector store when the repo changes
• Account system, authentication, rate-limit