Skip to content
gpuvox_
ConsoleRent a GPU

Quickstart

Four steps from this page to a root shell on a GPU machine.

  1. 1

    Find a machine

    Browse the marketplace or query the API. Sort by price per GB when memory is what you need.

    $ curl "https://gpuvox.com/api/v1/offers?min_vram=48&max_price=1"
  2. 2

    Rent it

    Create an account, add your SSH public key and credit, then press Rent and choose an image to start with. Without an account, the API below leaves a request that we answer by email.

    $ curl -X POST https://gpuvox.com/api/v1/rent-requests \
  -H "Content-Type: application/json" \
  -d '{"offerId":"vox-strix-01","email":"[email protected]",
       "sshKey":"ssh-ed25519 AAAA...","hours":24,"image":"llama-cpp"}'
  3. 3

    Wait a minute or two

    The machine starts right away. Its SSH command appears on your dashboard, and billing by the second begins.

  4. 4

    Connect

    The machine is yours until you release it on the dashboard. Billing stops the same second.

    $ ssh -p <port> root@<host from your dashboard>

AMD or NVIDIA

Partner machines are NVIDIA and run CUDA images. Our Strix Halo machine has an AMD Radeon 8060S: use llama.cpp built for Vulkan or ROCm, or PyTorch with ROCm.