More NVIDIA GPU stock now available!

Python client

Cudo Compute has a python client to manage all of your resources within the platform.

Install cudoctl

To use this client you will need to

  1. Create an API key from the web UI
  2. Download and install cudoctl command line tool from here
  3. Run cudoctl init and enter your API key and select a project/billing account.

Once cudoctl is set up the python client will use the default API key configured by cudoctl. If you would like to change project in the python client you will need to run cudoctl init again.

Install via pip

To install the python client via pip run:

pip install cudo-compute

Example code

For more information see the GitHub Repository Full examples can be found in the examples directory. Below are some simplified examples:

Create a VM

from cudo_compute import cudo_api, Disk, CreateVMBody

disk = Disk(storage_class="STORAGE_CLASS_NETWORK", size_gib=100,
                id="my-disk-id")

request = CreateVMBody(vm_id="my-vm-id", machine_type="epyc-rome-rtx-a4000",
                           data_center_id="no-luster-1", boot_disk_image_id='ubuntu-nvidia-docker',
                           memory_gib=16, vcpus=4, gpus=1, gpu_model="A4000", boot_disk=disk,
                           metadata={"group":"a"})

api = cudo_api.virtual_machines()
vm = api.create_vm(cudo_api.project_id(), request)

List current VMs

from cudo_compute import cudo_api

api = cudo_api.virtual_machines()
vms = api.list_vms(cudo_api.project_id())

Terminate a VM

from cudo_compute import cudo_api

api = cudo_api.virtual_machines()
api.terminate_vm(cudo_api.project_id(), 'my-vm-id')

Other services

The API has multiple services (other than virtual_machines)

  • apikeys
  • disks
  • networks
  • object_storage
  • permissions
  • ssh_keys
  • search
  • user
  • virtual_machines

Other services can be used like this:

from cudo_compute import cudo_api

api = cudo_api.networks()

For more information see the /docs directory of the GitHub Repository

DocsCLI Tool