Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.ornn.com/llms.txt

Use this file to discover all available pages before exploring further.

VM access is the default and recommended way to use your Ornn GPU reservation. Ornn provisions a managed virtual machine on your reserved hardware, pre-configured with your chosen image and your SSH public key. Once the VM is ready, your SSH endpoint and connection details appear on the reservation detail page in Portfolio.

​
Prerequisites

  • An accepted reservation in your portfolio
  • At least one SSH public key registered with your account
  • Access mode set to VM at /portfolio/access

​
Set up VM access

1

Go to /portfolio/access

Navigate to /portfolio/access. Use the Portfolio reservation dropdown to select the reservation you want to configure.
2

Select VM

Click the VM card. VM is selected by default on new reservations.
3

Choose a VM image

Under VM image, select the image you want Ornn to use when provisioning your machine:
  • Ornn base image · Ubuntu + CUDA + PyTorch — the default. A clean Ubuntu environment with NVIDIA drivers, CUDA, and PyTorch pre-installed. This is the right choice for most workloads.
  • Custom image — if you have approved custom images on your account, they appear in the dropdown alongside the base image.
The image selector shows a detail line describing the selected image.
4

Save your access settings

Your selection is saved as soon as you click the VM card. A confirmation message appears on the page. You can change the image selection and save again at any time before the reservation starts.
5

Connect via SSH

Navigate to your reservation detail page at /portfolio/[reservationId]. Once Ornn has provisioned the VM, the SSH endpoint and your tenant username appear on the page. Use them to connect:
ssh <tenant-username>@<ssh-endpoint>
Your registered SSH public key is automatically authorized on the VM.

​
The Ornn base image

The Ornn base image is an Ubuntu-based environment ready for GPU workloads out of the box:
  • OS: Ubuntu (22.04 or 24.04)
  • GPU drivers: NVIDIA drivers pre-installed and validated
  • CUDA: Installed and configured
  • Framework: PyTorch included
You do not need to configure drivers or install CUDA manually when using the base image.

​
Custom images

If you have uploaded your own VM image, it appears in the image dropdown once it has been reviewed and approved. The dropdown label shows the image display name and OS family; the detail line below it shows the image slug and a partial hash for verification.
Custom images must pass an Ornn security scan before they become available for selection. To submit an image for approval, contact Ornn support.
Only images with a clean scan status are shown in the dropdown. Images that are pending review, blocked, or revoked are not available for selection.

​
What’s next

Access overview

Compare VM and Bare Metal and understand when to use each mode.

Bare Metal access

Switch to direct host access for workloads that need maximum performance.