embedding models

A practical, Linux-first guide to stopping LLM hallucinations with Retrieval-Augmented Generation: prepare and chunk docs, embed with sentence-transformers into Chroma, retrieve wide then rerank with a cross-encoder, prompt defensively with citations and token budgets, and evaluate with ragas. Includes apt/dnf/zypper setup, end-to-end scripts, local LLM options, and tips for reproducibility, privacy, and automation.