Kain-90/RAG-Play

Provides step-by-step visualization of the entire RAG pipeline—from document chunking and vector embedding through semantic search to LLM response generation—using Transformers.js for client-side embeddings and LangChain for text processing. Built as a Next.js 14 frontend with Web Workers for efficient embedding computation, enabling real-time demonstrations without server dependencies. Includes UMAP-based dimensionality reduction for visualizing high-dimensional vector relationships and configurable retrieval strategies.