Set your own auto-stop timeout

Your GPU stops after a stretch of inactivity, so you are never billed for a machine you forgot about. That timeout used to be the same for everyone. Now it is yours to set.

• Set LLM and Media timeouts independently on the Settings page.
• Choose anything from 5 minutes up — or switch auto-stop off entirely for long jobs.
• Media treats a busy ComfyUI queue as activity, so a render that runs for hours keeps the pod alive. The countdown only starts once the queue is empty.
• LLM counts your own session, so your timeout never cuts anyone else off.

Running out of balance always stops a GPU, whatever you choose here.