Helm Control

Helm RuntimeInside Helm Node

From model
to running service.

Configure, start and observe your local inference runtimes. Keep launch settings, readiness and client access in view.

Explore Helm NodeSee what’s inside

Available within Helm Node.

A model package feeds a runtime, which connects to two API endpoints.

Saved profiles

Keep repeatable configurations for the engine, model, placement, context and concurrency.

Start and drain

Bring profiles online, inspect readiness and let active requests finish before stopping.

Logs and readiness

Understand runtime state through health checks, bounded logs and endpoint details.

CAPABILITIES, CLEARLY DEFINED

Built on a
practical foundation.

vLLM support requires a configured controller and a supported model, recipe and GPU combination.

Inside Helm Node today
  • llama.cpp and restricted vLLM recipes
  • Profile lifecycle, readiness and logs
  • Supported inference endpoints and usage history
PLANNED

General container and arbitrary service orchestration.

Product direction, not a release commitment.

PART OF SOMETHING CONNECTED

The Helm family
Helm FleetHelm NodeHelm ComputeHelm RuntimeHelm VaultPlannedHelm FlowPlanned