Service Status
Inference availability depends on the GPU server being online.
GPU Compute
Website & API
Operational
The GPU server is hosted on our own compute and may be offline during maintenance or power outages. Inference requests made while the GPU is offline return 503 and are not charged.