​
Prerequisites

  • A web browser
  • An SSH client (ssh on macOS/Linux, Windows Terminal, or PuTTY)
  • A credit or debit card, for when you add credits
1

Create an account

Go to platform.deploygpu.ai and sign up with Google, GitHub, an email magic link, or an email and password. No card is required to create the account.A default SSH key is generated for you at signup. Its private key can be downloaded once, for about 10 minutes after signup, so download it before you move on.
2

Add credits

Navigate to Billing in the sidebar.Click one of the quick-add buttons (10,10, 25, 50,50, 100, 250,250, 500) or enter a custom amount, and complete payment through Stripe. Every top-up carries a card processing fee of 5% plus $0.30, shown before you confirm.Your credit balance updates immediately after payment.
3

Choose a GPU

Go to Compute Marketplace in the sidebar.
  1. Select a GPU type from the cards (for example A6000 or A100).
  2. Optionally filter by region, number of GPUs, or interconnect type.
  3. The offers table shows live pricing sorted by hourly cost.
  4. Click Launch on the offer you want.
To start from an app instead, open Templates in the sidebar, choose the environment, then pick one of the suggested GPUs or browse all GPUs with the template kept.
4

Configure and deploy

On the deploy page:
  1. Template: pick PyTorch + Jupyter, ComfyUI, Open WebUI + Ollama, or the plain GPU VM, depending on the offer.
  2. Disk size: adjust if the offer allows it (some offers have a fixed size).
  3. SSH key: the key you used most recently is preselected, and your own keys come before the signup key, labelled (from signup). Make sure it is a key whose private half you have.
  4. Review the cost summary, which shows the offer’s rate and any platform fee already included.
  5. Click Deploy.
A plain GPU VM is typically running in 1 to 3 minutes. A template with an application, such as Open WebUI, takes about 6 to 8 minutes, because the platform waits for the application itself to start responding, not just for the machine to boot.
5

Connect, or open your app

Once your deployment shows Running on the Deployments page, you will see either a Copy SSH login control, an Open Application button, or both, depending on the template.For SSH:
Replace <port>, <user>, and <host> with the values shown on your deployment. The user depends on the provider: see Connect via SSH for the full table.For a template with an application (PyTorch + Jupyter, ComfyUI, or Open WebUI + Ollama), click Open Application instead. It opens a private link that only your DeployGPU account can use. There is no separate Jupyter URL shown next to the SSH details.
The dashboard shows Running as soon as the machine is up. SSH can take a little longer to start answering on some providers, so a connection refused right after Running usually clears within a minute.

​
When you are done

Stopping an instance terminates it and deletes its disk. There is no pause that stops billing. Copy off anything you want to keep first, then stop from the Deployments page: unused seconds in the current hour are refunded to your balance. See the stop instance guide for the full checklist, and Volumes if you want storage that survives between sessions.

​
Next steps

Templates

See what each of the four templates installs.

Volumes

Create persistent storage that survives across deployments.

API Keys

Generate API keys for programmatic access.

Billing

Set up auto top-up so you never run out of credits.