Registrieren

Reliability-aware LLM router: run requests on your local model and automatically fall back to a remote model when local inference stalls or crashes.