Compute
GPU and CPU availability for workload planning.
A frontend-only availability surface for selecting compute by region, estimated latency, cost, memory, and current capacity once backend data is connected.
Availability console
Select compute before the workload runs.
The interface is designed for live infrastructure data later. It intentionally contains no static instance names, prices, latency numbers, availability counts, or region inventory.
This page is frontend-only. It does not ship with static instance names, pricing, latency, region, or availability data. Those fields will populate from backend services later.
Filter GPUs by VRAM
Backend range pending
No live GPU availability is connected yet.
The backend should provide instance identity, region, latency, hourly cost, memory, CPU/GPU capacity, and availability status. Until then, this surface stays empty by design.
Region-aware
Region options should be supplied by the backend, not hardcoded in the UI.
Latency-aware
Latency should reflect live or recently measured backend data for the selected workload context.
Cost-aware
Cost and availability should arrive from the pricing and capacity layer when it is connected.