CLI reference

The llmkube CLI is a thin client over the Kubernetes API: it knows how to construct Model and InferenceService manifests from a small set of flags.

What this page will cover

llmkube deploy: deploy a catalog model or a custom GGUF source URL.
llmkube catalog list / info: browse the bundled model catalog.
llmkube status: check phase, replicas, and endpoint for a deployed InferenceService.
llmkube version: print the client build and check for a newer release.
The other top-level commands: list, delete, scale, queue, benchmark, cache, inspect, license, audit, foreman, fleet, and metal.