TrackMyGPU, a market radar for cloud GPUs

Field notes

The service organizes cloud GPU selection around current hourly instance offers rather than generic hardware specifications. Its advisor asks users to describe a model or workload, budget, duration, and reliability requirement, then maps that request to VRAM, hardware, provider, and price options. The workspace adds saved offers, price targets, encrypted RunPod or Vast.ai credentials, deployment monitoring, and an auto-stop period intended to limit billable runtime.