This ComfyUI workflow is designed to generate a 3D model from a 2D image, while also containing a separate text-to-image generation process.

The primary function begins by loading an input image. The image is then resized to 518x518 pixels and its background is automatically removed to create a mask of the main subject.

This processed image and its mask are fed into a Hunyuan3D model. The workflow utilizes this model to generate a 3D mesh, which is then decoded and post-processed to refine its quality. The final output is a 3D model exported as a .glb file, which can be viewed in a 3D previewer node within the interface.

Additionally, the workflow includes a standard text-to-image pipeline using the Juggernaut XL checkpoint. This separate section uses text prompts to generate a 2D image, which is saved independently from the 3D model creation process.

download hunyuan 3d and put it in models/diffusion_models\hy3dgen

and juggernautXL_ragnarokBy and out in models\checkpoints in your comfy ui folder

note: it doesnt have the texture part as it have problems, but i am working on a new custom node