ComfyUI 3D Guide

Rated Excellent/5

Open Source & Free

ComfyUI 3D Model Generation Workflow - Node-based interface for Hunyuan3D

Complete Tutorial

How to Generate 3D Models with ComfyUI

Learn how to create high-quality 3D models using ComfyUI and Hunyuan3D-2. Step-by-step guide for image-to-3D generation with node-based workflows.

Free

Open Source

12GB

Min VRAM

GLB

Output Format

Want to skip the setup?

Generate 3D models instantly on 3D AI Studio with Trellis 2 or Hunyuan 3D - no installation, no GPU required.

Introduction

What is ComfyUI?

ComfyUI is a powerful, open-source graphical user interface for running AI models locally on your computer. Unlike traditional software, ComfyUI uses a node-based workflow system where you visually connect different AI components to create complex generation pipelines.

For 3D model generation, ComfyUI integrates with models like Hunyuan3D-2.0 from Tencent, enabling you to convert images into high-quality 3D meshes with textures directly on your hardware.

The node-based approach gives you full control over the generation process. You can customize every step, combine multiple AI models, and create reusable workflows for consistent results.

100% Free & Open Source

Runs on Your Hardware

Full Privacy Control

Highly Customizable

ComfyUI Node-Based Interface - Visual workflow editor for AI models

Why ComfyUI for 3D

ComfyUI 3D Generation Features

A powerful open-source platform for AI-powered 3D model creation

Node-Based Workflow

Build complex 3D generation pipelines with visual nodes. Connect different AI models and processing steps.

100+custom nodes

Multiple 3D Models

Access Hunyuan3D-2, Hunyuan3D-2mv, and turbo variants. Generate 3D from single or multi-view images.

6+3D models

Local Processing

Run everything on your own hardware. No cloud costs, full privacy, and offline capability.

12GBmin VRAM

Open Source

Free and open-source software. Active community with regular updates and new features.

Freeforever

Hunyuan3D 2.0 in ComfyUI

Hunyuan3D 2.0 is Tencent's open-source 3D asset generation model, capable of generating high-fidelity 3D models with high-resolution texture maps from text or images.

Geometry Generation (Hunyuan3D-DiT)

Based on a flow diffusion Transformer architecture, this model generates untextured geometric models that precisely match input conditions. It creates the shape and structure of your 3D object.

Flow DiffusionTransformerClean Geometry

Texture Generation (Hunyuan3D-Paint)

Combines geometric conditions and multi-view diffusion techniques to add high-resolution textures to models. Supports PBR materials for realistic lighting effects.

Multi-View DiffusionPBR MaterialsHigh-Res Textures

Key Advantages of Hunyuan3D 2.0

High-Precision Generation

Sharp geometric structures, rich texture colors, PBR material support for realistic lighting.

Diverse Usage Methods

Code calls, Blender plugins, Gradio apps, and ComfyUI integration for different workflows.

Lightweight & Compatible

Mini model needs only 5GB VRAM. Full pipeline (shape + texture) requires just 12GB.

Step-by-Step Guide

How to Install ComfyUI for 3D Generation

Follow these steps to set up ComfyUI on your local machine for 3D model generation

Prerequisites

Miniconda or Anaconda installed for Python environment management
Git installed on your system
Microsoft Visual C++ Redistributable installed
NVIDIA GPU with CUDA support (recommended) or AMD GPU

Create Virtual Environment

Set up an isolated Python environment using Conda to avoid dependency conflicts.

conda create -n comfyenv
conda activate comfyenv

Clone ComfyUI Repository

Download the ComfyUI source code from GitHub to your local machine.

git clone git@github.com:comfyanonymous/ComfyUI.git

Install Dependencies

Install all required Python packages including PyTorch and GPU drivers.

cd ComfyUI
pip install -r requirements.txt

Start ComfyUI

Launch the application and access it through your web browser.

python main.py

GPU Dependencies

NVIDIA (Recommended)

Install CUDA Toolkit and cuDNN. PyTorch will detect automatically.

AMD

Install ROCm for Linux. Windows support is limited.

Mac ARM Silicon

MPS backend is supported. Performance varies by model.

ComfyUI 3D Workflows

Hunyuan3D-2 Workflow Examples

Choose the right workflow based on your input images and requirements

Hunyuan3D-2mv

Multi-View Input

Generate 3D models using multiple view images (front, side, back). Best for complex objects where different angles help define the shape accurately.

Multiple angle inputs
Higher accuracy
Complex geometry support

6GB+ VRAM

VRAM

~2-3 min

Gen Time

GLB

Output

Steps to Run the Workflow

1
Download the workflow image from the ComfyUI documentation and drag it into ComfyUI to load the workflow.
2
Ensure the model is loaded in the Image Only Checkpoint Loader node:
hunyuan3d-dit-v2-mv.safetensors
3
Load your input image(s) in the Load Image node(s). Add front, side, and back views for best results.
4
Click Queue or press Ctrl+Enter to run the workflow.
✓
Find your model in the ComfyUI/output/mesh folder.

Model file location:

ComfyUI/models/checkpoints/hunyuan3d-dit-v2-mv.safetensors

ComfyUI Hunyuan3D-2 Workflow Example - Image to 3D model generation

Drag workflow PNG images into ComfyUI to load workflows automatically

Model Downloads

Hunyuan3D 2.0 Models for ComfyUI

Download the model files and place them in your ComfyUI/models/checkpoints folder

Model	Description	Parameters	Date
Hunyuan3D-DiT-v2-mv `hunyuan3d-dit-v2-mv.safetensors`	Multi-view image to shape model	1.1B	March 2025
Hunyuan3D-DiT-v2-mv-turbo `hunyuan3d-dit-v2-mv-turbo.safetensors`	Fast distilled multi-view model	1.1B	March 2025
Hunyuan3D-DiT-v2-0 `hunyuan3d-dit-v2.safetensors`	Single image to shape model	1.1B	January 2025
Hunyuan3D-DiT-v2-mini `hunyuan3d-dit-v2-mini.safetensors`	Lightweight mini model	0.6B	March 2025
Hunyuan3D-Paint-v2-0 `hunyuan3d-paint-v2.safetensors`	Texture generation model	1.3B	January 2025

Model Installation Path

ComfyUI/
├── models/
│   ├── checkpoints/
│   │   ├── hunyuan3d-dit-v2-mv.safetensors
│   │   ├── hunyuan3d-dit-v2-mv-turbo.safetensors
│   │   ├── hunyuan3d-dit-v2.safetensors
│   │   └── hunyuan3d-paint-v2.safetensors

Download models from Hugging Face and rename them to the filenames shown above before placing in the checkpoints folder.

Hardware Requirements

System Requirements for ComfyUI 3D

Make sure your hardware meets these requirements for smooth 3D generation

Minimum Requirements

GPU VRAM

6GB (shape only)

Full Pipeline

12GB VRAM

Mini Model

5GB VRAM

RAM

16GB+

ComfyUI vs Web-Based 3D Generation

Choose the right approach for your 3D model generation needs

ComfyUI (Local)

Full control, local processing

Setup Time

30-60 minutes

Technical Knowledge

Required

GPU Required

Yes (12GB+)

Cost

Free (hardware costs)

Privacy

Full control

Customization

Extensive

Best for: Advanced users who want full control, have powerful hardware, and need extensive customization.

Easiest Option

3D AI Studio (Web)

Instant access, no setup

Setup Time

None - instant

Technical Knowledge

Not needed

GPU Required

Cost

Pay per use

Privacy

Cloud-based

Customization

Preset options

Best for: Users who want quick results without setup, don't have a powerful GPU, or need to generate 3D models on the go.

How to Update ComfyUI

Keep your ComfyUI installation up to date for the latest features and bug fixes

Pull the latest code

Navigate to your ComfyUI installation directory and pull updates from GitHub.

cd <installation path>/ComfyUI
git pull

Update dependencies

Ensure your Conda environment is activated, then reinstall dependencies.

conda activate comfyenv
pip install -r requirements.txt

Important: Make sure to update regularly as native Hunyuan3D-2mv support was added recently. Older versions may not include 3D generation features.

FAQ