I Built a RAG System That Runs Completely Offline

Most RAG tutorials follow a familiar pattern: Upload documents → send them to a cloud service → generate embeddings → retrieve context → ask an LLM. But I wanted to try something different: Can we bui