**Describe the feature** Could we somehow integrate LLMs into Kool? Docker released a feature to run LLMs inside containers. Not sure if Kool would make a nice use of it: https://docs.docker.com/desktop/features/model-runner/