Model routing

Local when you want local. Cloud only when you choose it.

Burnout Bytes should make the route visible. A user can run local models such as Gemma or Qwen for private work, and can explicitly choose a cloud model when a task needs that capability.