Ollama CLI

The ollama command line is a client of whatever OLLAMA_HOST names, and the server answers the Ollama API.

# install: https://ollama.com/download
superfluid launch ollama --model unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_M               # ollama run <model>
superfluid launch ollama -- list
superfluid launch ollama -- show unsloth/Qwen3.8-27B-GGUF

What launch gives it

given	what for
OLLAMA_HOST	the server's address
run <model>	the default command; arguments after -- replace it

Model names are the server's model ids; ollama pull and Modelfiles are not served (see Ollama API).