🔨 I have prompts working (but not streaming) with llm-ollama. It needs a lot of work and polish still.