REVIEW PAPER ON AGENTIC AI BASE E LEARNING SYSTEM

– Traditional e-learning tools often deliver fixed content that fails to adapt to individual pace or offline access, particularly when studying from uploaded books. The proposed system introduces an agentic AI core that automatically processes PDF/TXT books into structured chapters and topics, stores embeddings in a FAISS vector database, and enables interactive learning through RAG-based chat, on-the-fly MCQ generation, and instant evaluation. Running entirely on a local Flask server with Ollama LLM and SQLite storage, it supports seamless flows from book upload to performance dashboards without internet dependency. Modular components—BookProcessor, MCQGenerator, TestEngine, and ChatRAG—ensure easy maintenance and scalability.