🧠 Notebook Colleague is a full-stack, AI-powered research assistant inspired by Google's NotebookLM. It allows users to create isolated workspaces ("Labs"), upload PDF documents, and interact with them using Retrieval-Augmented Generation (RAG). It features a highly interactive UI with a neural network background animation, resizable panels, and integrated web search.

✨ Features 🧪 Multi-Lab Architecture: Create, rename, and delete multiple "Notebooks" (Labs) to keep different research projects isolated. 📄 RAG Document Chat: Upload up to 10 PDFs per lab. The AI (Llama-3 via Groq) answers questions strictly based on the selected documents. 🌍 Quick Web Chat: A dedicated right-hand sidebar connected to DuckDuckGo Search for finding live information outside your documents. 🎨 Dynamic UI: Neural Network Background: Interactive HTML5 Canvas animation with nodes connecting based on proximity. Resizable Layout: Drag-and-drop handler to adjust the width of the chat vs. web search panels. ⚡ High Performance: Uses Groq API for near-instant inference and ChromaDB for local vector storage.

NOTE: Use your own Groq API for the project.

🛠️ Tech Stack Backend: Framework: Flask (Python) LLM: Llama-3-70b-versatile (via Groq) Vector Store: ChromaDB (Persisted locally) Embeddings: HuggingFace (all-MiniLM-L6-v2) Orchestration: LangChain Search Tool: DuckDuckGo Search

Frontend: Core: Vanilla HTML5, CSS3, JavaScript (ES6+). Visualization: HTML5 Canvas (Custom particle system). Styling: CSS Variables, Flexbox/Grid, Backdrop Filters.

Created by G.S.Ganesh Subramanian.