Frequently asked questions

Billing starts the moment your pod reaches the RUNNING state and stops the second you stop or terminate it — no rounding up to the hour and no minimum runtime. Charges are drawn from your prepaid balance and itemized in your dashboard in real time.

We offer NVIDIA H100 (80 GB), A100 (80 GB), RTX 4090 (24 GB), and RTX 3090 (24 GB), in both on-demand and interruptible tiers. See the full rate card on the pricing page.

On-demand pods provide dedicated, non-interruptible capacity that runs until you stop it. Interruptible pods use spare capacity at up to 60% lower cost but may be reclaimed during peak demand, with a 30-second graceful-shutdown warning. Interruptible is ideal for fault-tolerant and batch jobs.

Create a pod via the CLI or dashboard, connect over SSH, and run your container or framework. Our documentation includes a step-by-step guide for serving an LLM with vLLM in a few commands.

All connections use TLS 1.2+, pods are network-isolated, and SSH uses key-based authentication only. We do not inspect the contents of your workloads except where strictly necessary to investigate a policy violation or comply with a valid legal request. See our Privacy Policy.

You add credit to a prepaid balance, and charges are metered per second as you use resources. Unused balance is fully refundable within 14 days of deposit, and you can request a refund of any remaining balance at any time. Full terms are in our Refund Policy.

Cryptocurrency mining, malware hosting, phishing, DDoS attacks, and other illegal or abusive activities are strictly prohibited and result in an immediate account ban without refund. Read the full Acceptable Use Policy.

Yes. Reach our engineers at [email protected] or through the contact form. We typically reply within one business day.