Frequently asked questions

Per second while a workload runs, from a prepaid wallet you top up by card. There's no subscription and no minimum. Charges are in US dollars; amounts are also shown in CA$ for convenience.

We email you when your balance runs low. If it reaches zero, running workloads are stopped so you never spend more than you have. Storage volumes are kept, so your data is safe; add funds and deploy again.

Yes, for workloads with a storage volume. While paused, GPU charges stop and only the storage is billed. Resume when you need it again.

In our provider's data centers. Each GPU's regions are listed on the pricing page, and you see the region when you deploy.

Any model from our catalogue, or any Hugging Face model ID, served behind an OpenAI-compatible API. Gated models need you to accept their licence on Hugging Face and add your Hugging Face token.

No. Models run inside your own workload, and we don't train any model on your files, prompts or outputs.

Unused card top-ups can be refunded to the card that paid them from your Wallet, once nothing is running. Coupon credit can't be refunded. If a deploy fails on our side, that time is refunded automatically.

Not a formal SLA today. You can check live status anytime on our status page, and failed deploys are refunded automatically. Contact us if your organization needs contractual commitments.

Yes. A team has its own wallet, workloads and storage. One owner manages billing; admins manage members and use everything else; members deploy and use the team's resources.

A private AI workspace that runs on its own GPU. Connect Dropbox, Google Drive, OneDrive, S3 or SFTP, and ask questions about your documents or generate Word and Excel files — your files stay in your workload.

Open a ticket from Support in the console, in English or French, or use the contact page.