External power on demand
Load spikes, demos, or a one-off high-cost job: route to OpenRouter without redesigning your stack. The panel stays the control point.
Extension · OpenRouter
Xeretron is integrated with OpenRouter to use external power when you need it and giant models when they are truly required and worth it. Local remains the default; external is a deliberate exception.
Why it exists
Junior, Developer, Turbo, Titan, and Omega cover day-to-day on your GPU. OpenRouter steps in when you need more power than fits in the rack — or a Frontier model that does not make sense to host permanently.
Load spikes, demos, or a one-off high-cost job: route to OpenRouter without redesigning your stack. The panel stays the control point.
When the case truly requires it and the ROI justifies it — not for fashion. Big frontiers for big problems; everything else returns to the node.
One bridge to many providers and models. You set credentials and policy; Xeretron does not force you to live in the cloud.
Privacy, predictable cost, and rack latency. External is an extension — not the product’s center of gravity.
If the work fits Titan or Omega locally, it stays there. OpenRouter for what your hardware cannot reach or does not justify buying.
Every external hop means data leaving the node. Use it with clear policy: what leaves, what does not, and who authorizes the provider.
How it fits
You work with the Xeretron catalog and HF locally. When an agent, bench, or delivery needs a model that does not fit (or should not live) on your GPU, you choose the OpenRouter path from the same ecosystem.
Next step
We evaluate with you when OpenRouter adds value — and when your GPU already covers it.