NotebookLLM — RAG-Powered Document Chat

A Google NotebookLM clone built with Next.js that allows users to upload documents (PDF/CSV) and have intelligent conversations with them using Retrieval-Augmented Generation (RAG).

🎯 Project Overview This application implements a complete RAG pipeline where users can:

• Upload PDF or CSV documents
• Have the system intelligently chunk, embed, and index the content
• Ask natural language questions about the document
• Receive grounded answers based solely on the document's content (no hallucinations)

Built as part of Assignment 03 — Google NotebookLM RAG to demonstrate end-to-end RAG implementation.

🏗️ Architecture

• LLM: Groq (Llama 3.1)
• Embeddings: Hugging Face Inference API (BAAI/bge-small-en-v1.5)
• Vector Database: Qdrant Cloud
• Framework: Next.js 16 (App Router)
• RAG Orchestration: LangChain

🚀 Getting Started

1. Clone the repository
2. Install dependencies:

   npm install --legacy-peer-deps

3. Setup Environment Variables: Create a .env file in the root directory:

   HF_TOKEN=your_huggingface_token
   GROQ_API_KEY=your_groq_api_key
   QDRANT_URL=your_qdrant_url
   QDRANT_API_KEY=your_qdrant_api_key

4. Run the application:

   npm run dev

🛠️ Tech Stack

• Frontend: React 19, Tailwind CSS, Framer Motion, Lucide React
• Backend: Next.js API Routes (Node.js)
• AI/ML: LangChain, Groq API, Hugging Face Inference API
• Storage: Qdrant Vector DB