txt2glb

Changes:
Workflow completely revised and simplified.

3D/GLP:
Now uses the Simplified Subgraph workflow from hunyuan3D to simplify the entire flow.

Image:
Keep the object simple and the prompt short.

PLAIN WHITE BACKGROUND!!! LowPoly works best for me.
The 1:1 ratio delivered the best results. 512x512 or 784x784 (recommended)

My first Comfy workflow.

More Just4Fun than a serious workflow.
Your prompt generates an image, which is then converted into a 3D object in GLB format.

There are probably better ways to do this, but it's quite useful for experimenting with different checkpoints for the images.

Uses hunyuan3d-dit-v2_fp16.safetensors for 3D generation.

Tip: Keep the objects simple, not too complex.

Sort by

No Images Found.