Dev: ComfyUI Integration

Development Update: Prism ComfyUI Integration

Today we want to share the first teaser of our upcoming ComfyUI integration with you. This will allow studios to integrate new AI workflows seamlessly into their existing studio pipelines.

We showed a preview of the ComfyUI integration last month at FMX – Film & Media Exchange.
The full recording will be available soon.

Here’s a summary of the features of the ComfyUI plugin:

• save, load, and version your ComfyUI workflows using Prism
• load images, videos, and 3D assets from other DCCs into ComfyUI
• export generated images and videos back into the project to bring them into other DCCs
• run ComfyUI workflows on Deadline
• integrate with the Prism Execution Graph to automate the triggering of ComfyUI workflows
• run workflows locally or in the ComfyUI Cloud

In the video, we also show an interesting use case of the ComfyUI plugin:

After rendering a few shots in Houdini, we use the rendered images as references for a new shot. In ComfyUI, we apply the lighting and materials from the existing shots to a playblast of the new shot.
The ComfyUI workflow is triggered automatically after creating a playblast in Houdini and is submitted as a Deadline job.
The resulting video can give us an idea of how the shot will look in the sequence before rendering a single frame of it in Houdini.
This is just one example, but with the vast amount of custom nodes available in ComfyUI, there are no limits to what you can create.

The ComfyUI Prism plugin will be available in Early Access soon.