Your workflow.
Anywhere.
Run ComfyUI on a rented GPU using our ready cloud template.
A prepared starting point
The public ComfyUI template includes nodes, dependencies and model folders. You can use it with your own workflows; buying a pack is not required.
This is our Seedra partner link. Vast.ai may pay us a commission; your price stays the same.
Choose enough GPU memory
| Work | Starting point |
|---|---|
| Photo workflows | 8 GB+ VRAM |
| Video workflows | 32 GB+ VRAM |
Exact requirements depend on the pack. Compare the current offers in Vast.ai before renting. GPU rental is separate from product prices; live rates and availability change.
From template to first result
Start your instance
Open the template and choose a GPU that meets your workflow requirements.
Prepare the models
Run the model install command included with your pack. That installer expects this template’s folder structure.
Load and test
Open the workflow .json in ComfyUI and run a small test before a longer session.
When finished, stop the instance and review any remaining storage charges in the provider dashboard.
Cloud launch and studio preview
Cloud launch takes you to Vast.ai. The SEEDRAAI studio is a separate preview with prepared samples. Opening this page does not rent a GPU or start a paid instance.