ComfyUI Extension: ComfyUI-Hunyuan3D-v3

Authored by exedesign

Created

Updated

1 stars

Text-to-3D and Image-to-3D generation using Tencent Cloud Hunyuan 3D Global API. Supports PBR materials, face count control (40K-1.5M faces), and multiple generation types (Normal/LowPoly/Geometry/Sketch). Outputs industry-standard GLB format. Requires Tencent Cloud account with API access.

Custom Nodes (0)

    README

    ComfyUI-Hunyuan3D-v3

    License: MIT GitHub release

    ComfyUI custom node for Text-to-3D and Image-to-3D generation using Tencent Cloud Hunyuan 3D Global API v3.

    Features

    • Text-to-3D: Create 3D models from text prompts
    • Image-to-3D: Convert images to 3D models
    • Advanced Options: PBR materials, face count control (40K-1.5M), generation types
    • GLB Output: Industry-standard format
    • Global API: ap-singapore region

    Installation

    Via ComfyUI Manager

    1. Open ComfyUI Manager
    2. Search "Hunyuan 3D"
    3. Install & Restart

    Manual

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

    Setup

    1. Get API keys: https://console.intl.cloud.tencent.com/ Access Management API Keys
    2. Activate Hunyuan 3D service (Console AI Services)
    3. Add billing info ( Paid API: ~$0.10-0.60 per request)

    Quick Start

    HunyuanConfig (SecretId, SecretKey)  HunyuanTextTo3D  Preview3D

    Parameters

    • enable_pbr: PBR materials (default: False)
    • face_count: 40000-1500000 (default: 500000)
    • generate_type: Normal/LowPoly/Geometry/Sketch
    • polygon_type: triangle/quadrilateral

    Output

    • Format: GLB
    • Location: ComfyUI/models/3d_models/

    Batch Processing

    ⚡ Method 3: Automatic Batch Node (EASIEST!)

    One-click batch processing for entire folders!

    1. Load: examples/auto_batch_workflow.json
2. Put images in ComfyUI/input/
3. Set API credentials
4. Configure input/output folders
5. Click Queue Prompt once → ALL images processed!

    Features:

    • ✅ Single click processes all images
    • ✅ Automatic file discovery with pattern matching
    • ✅ Custom output folder selection
    • ✅ Progress tracking per image
    • ✅ Detailed summary report with paths

    Method 1: Python Script

    # 1. Copy batch_hunyuan3d.py to ComfyUI root folder
# 2. Edit API credentials in the script
# 3. Place images in 'input_images' folder
# 4. Run:
python batch_hunyuan3d.py

    Method 2: Manual Workflow

    Use examples/batch_image_to_3d_workflow.json:

    1. Load workflow in ComfyUI
    2. Select image and Queue Prompt
    3. Repeat for each image

    Troubleshooting

    ResourceInsufficient Error:

    • Check account balance
    • Add credits ($10+ recommended)
    • Verify service activated

    Batch Processing Tips:

    • Process 5-10 images at a time
    • Each image takes ~2-5 minutes
    • Monitor API credits

    License

    MIT - See LICENSE

    Requires Tencent Cloud account with billing enabled