A self-hosted Kubernetes operator that analyzes your HPA and KEDA autoscaling with a rule engine plus an LLM on your own key, then recommends - and optionally applies - better config. Runs entirely in your cluster.