Hunyuan 3D Model Preview

What it is

Hunyuan 3D Model Preview is the viewer half of the ComfyUI-Hunyuan3D-v3 pack. Hunyuan Image to 3D and Hunyuan Text to 3D both hand you back a file path, not an image you can look at directly in the graph - this node is what turns that path into something you can actually inspect inside ComfyUI, instead of alt-tabbing to a file browser to check whether your generation came out right.

It's a small, single-purpose node, and that's the point: the generator nodes do the expensive API call, this one just loads and displays the result.

How it works

Give it the file path a generator node produced and it locates and loads that GLB so it renders inline in ComfyUI. There's no processing here - no re-export, no format conversion - it's purely a load-and-show step, which is why it's a separate node from the generators rather than baked into them: you might generate once and preview several times, or generate a batch and only want to check one result without re-running anything.

Inputs and outputs

The whole node is one input and one output:

• model_path - a string, the GLB file path. In practice this is almost always the model_path output of a Hunyuan Image to 3D or Hunyuan Text to 3D node wired straight in, though nothing stops you from typing a path by hand if you've got a GLB sitting on disk from somewhere else.

Output is model_path again - the same string passed straight through. That's a deliberate pattern in this pack: it lets you chain a preview node into something else downstream that also expects a path (or just re-display it further along the graph) without breaking the connection.

Install

Via ComfyUI Manager: search "Hunyuan 3D", install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/exedesign/Hunyuan-3D-v3.git
cd Hunyuan-3D-v3
pip install -r requirements.txt

This node has no API dependency of its own - no Hunyuan Config, no Tencent Cloud credentials needed, since it never talks to the network. It just reads a local file. That makes it the one node in this pack you can test in isolation: point it at any GLB you already have lying around to confirm the install worked, before you bother setting up billing on the Tencent side.

Troubleshooting

If nothing shows up, the most common cause is simply that model_path is empty or pointing at a file that doesn't exist yet - check that the upstream generator node actually finished successfully rather than erroring out or timing out, since a failed generation won't produce a valid path for this node to load. Beyond that, there isn't much surface area for this node to go wrong on its own; most problems that look like a preview issue actually trace back to the generation step that fed it.