Images

The Images API lets you list available operating system images that come pre-configured with CUDA drivers and GPU toolkits. You select an image when deploying a GPU cluster — it determines the base OS, driver stack, and pre-installed software.

Base URL: https://gpu.huddleapis.com/api/v1

What's in an Image?

Each image is a pre-built OS snapshot that includes:

Operating system (e.g., Ubuntu 22.04, Ubuntu 24.04)
CUDA drivers (e.g., CUDA 12.4, CUDA 13.0)
GPU toolkits and libraries
Python and common ML dependencies

Images are maintained and updated by Huddle to ensure driver compatibility with the underlying hardware.

List Available Images

curl -H "X-API-Key: mk_your_key" \
  https://gpu.huddleapis.com/api/v1/images

Response:

{
  "code": "OK",
  "data": [
    {
      "id": "e79a91e9-8750-44cb-af19-bf6ab83121ee",
      "name": "ubuntu-22.04-cuda-12.4-cluster",
      "image_type": "cluster",
      "details": [
        "Ubuntu 22.04",
        "CUDA 12.4",
        "Python 3.10"
      ]
    },
    {
      "id": "a3b2c1d0-1234-5678-9abc-def012345678",
      "name": "ubuntu-24.04-cuda-13.0-cluster",
      "image_type": "cluster",
      "details": [
        "Ubuntu 24.04",
        "CUDA 13.0",
        "Python 3.12"
      ]
    }
  ]
}

Filtering Images

Parameter	Type	Description	Default
`image_type`	string	Filter by image type (`cluster`, `spot`)	`cluster`
`cluster_type`	string	Filter to images compatible with a specific cluster type	—

# List images compatible with a specific GPU type
curl -H "X-API-Key: mk_your_key" \
  "https://gpu.huddleapis.com/api/v1/images?cluster_type=1H100.80S.30V"

Check compatibility before deploying

Not all images work on all GPU types. Always filter by cluster_type to see which images are compatible with the instance you plan to deploy.

Using an Image in a Deployment

Pass the image name (not ID) when creating a deployment:

curl -X POST -H "X-API-Key: mk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "cluster_type": "1H100.80S.30V",
    "image": "ubuntu-22.04-cuda-12.4-cluster",
    "hostname": "my-gpu",
    "ssh_key_ids": ["key-id"],
    "location": "FIN-01"
  }' \
  https://gpu.huddleapis.com/api/v1/deployments/clusters

Next Steps

Browse Offers to find GPU types to deploy
Learn the full Deployment workflow
Upload SSH Keys you'll need to access your instance