A retrieval-augmented generation (RAG) app. You upload PDFs, they're split, embedded and stored in a ChromaDB vector database, and the assistant answers questions in plain language with the exact file and page as its source. A LlamaIndex document agent adds data extraction, summaries and document comparison, all behind a FastAPI service. Built because plain chatbots often invent answers without a source. Next step: adding a reranker for better retrieval. Tools: Python, LangChain, LlamaIndex, ChromaDB, OpenAI API, FastAPI.