From zero to
GPU workload.

Start with the workload. Determine the model or application memory requirement, then compare GPU model, VRAM, GPU count, region and customer price in the deployment catalog.

• Choose enough VRAM with runtime headroom
• Select the region and configuration you need
• Review customer pricing before deployment

Create the workload from BurstDock and follow its lifecycle state. Once running, use the workspace to keep compute and usage visible.

For AI-agent workflows, use the BurstDock integration reference for setup, access scopes and supported operations.

Stop compute when appropriate and remove resources you no longer need. Review the product state and billing information shown in the workspace because storage or other resource behavior can differ from active compute.