Creating a preset

Click Create preset on the Presets list page. The wizard has three steps.

Step 1 — Identity

Field	Required	What it does
Display name	required	Human-readable name shown on tiles and in the session wizard.
Description	optional	Short explanation of what this preset is for.
IDE type	required	jupyterlab, vscode, or rstudio. Determines the icon and the default image.
Home PVC	optional	Default home volume for sessions created from this preset. If omitted, sessions default to creating a new 5 GiB PVC.

Step 2 — Pod sizes

Add one or more compute options. Each pod size defines the CPU, memory, and GPU available to a session.

Field	Required	What it does
Display name	required	Label shown in the session wizard's compute step (e.g. "Small", "1× GPU").
CPU	required	Number of CPUs or millicores (e.g. 2, 4000m).
Memory	required	Memory quantity (e.g. 8Gi, 32Gi).
GPU count	optional	Number of GPUs. Zero means CPU-only.
GPU vendor	optional	nvidia.com/gpu, amd.com/gpu, or habana.ai/gaudi.

One pod size must be marked as the default — it will be pre-selected in the session wizard.

Step 3 — Storage

Optionally pre-mount data volumes that every session created from this preset will inherit.

Field	Required	What it does
Data PVCs	optional	Persistent volume claims to mount automatically. Each entry specifies the claim name, mount path, and read-only flag.

Sessions can always add extra data volumes at creation time — these are just the ones baked into the preset.

After creation

The new preset appears on the list immediately. Click it to open the detail view, where you can:

• Browse the full image catalog and pod sizes.
• See which sessions are currently using this preset.
• Click Launch session to start a new session pre-filled with this preset.

tip

Presets are templates, not instances. Editing a preset changes what new sessions see, but it does not modify sessions that are already running.