Run a Full RAG Agent Offline with LangGraph, Ollama, and Embedded Qdrant

A developer demonstrates running a complete RAG agent locally using Ollama for chat and embeddings, and an embedded Qdrant vector store. The setup requires no API keys or Docker, just two Ollama models and a config flip. The approach leverages a provider-swap design to switch between local and cloud backends via configuration.