Last released Oct 5, 2026
Share one GPU between LLM servers and ComfyUI: a queue that swaps models in and out on demand.