Three commands. A local LLM running on your laptop, with an OpenAI-compatible HTTP API ready to call from anywhere.