rag_system/
├── documents/          ← Put your PDFs and Word files here
├── chroma_db/          ← Created automatically; Chroma persists here
├── config.py           ← Settings: chunk size, model names, thresholds
├── indexer.py          ← Ingest, chunk, embed, store (run once)
├── retriever.py        ← Retrieve, augment, generate (run per query)
├── rag_system.py       ← Top-level class tying it all together
└── main.py             ← Entry point: index documents, then answer questions
