ComfyUI-Meshy

ComfyUI-Meshy

ComfyUI-Meshy is a 3D generation extension based on Meshy AI API. It provides Text-to-3D and Image-to-3D capabilities within ComfyUI, allowing you to generate 3D models directly from text prompts or images.

Features

• Text to 3D (Preview): Generate preview 3D models from text descriptions
• Text to 3D (Refine): Refine preview models with textures and details
• Image to 3D: Convert images into 3D models
• Multiple AI Models: Support for Meshy-4, Meshy-5, and Meshy-6
• Customizable Output: Control topology (triangle/quad), polygon count, and PBR settings
• Multiple Art Styles: realistic, cartoon, low-poly, sculpture, pbr

Installation

Method 1: Manual Installation

1. Clone this repository to your ComfyUI custom_nodes directory:

cd ComfyUI/custom_nodes
git clone https://github.com/YOUR_USERNAME/ComfyUI-Meshy.git

2. Install dependencies:

pip install -r ComfyUI-Meshy/requirements.txt

3. Restart ComfyUI

Method 2: ComfyUI-Manager

Search for "ComfyUI-Meshy" in ComfyUI-Manager and click Install.

Getting Started

1. Get Your API Key

1. Visit Meshy AI
2. Create an account or sign in
3. Go to API settings and generate an API key

2. Basic Usage

1. Add "Meshy - API Key" node and enter your API key
2. Add "Meshy - Text to 3D (Preview)" or "Meshy - Image to 3D" node
3. Connect the API key to the generation node
4. Configure your parameters and run!

Nodes

Meshy - API Key

Input node for your Meshy API key. All generation nodes require this connection.

Meshy - Text to 3D (Preview)

Generate a preview 3D model from text description.

Parameters:

• prompt: Text description of the 3D model
• art_style: Visual style (realistic, cartoon, low-poly, sculpture, pbr)
• ai_model: Model version (meshy-4, meshy-5, meshy-6)
• topology: Mesh topology (triangle, quad)
• target_polycount: Target polygon count (100-300,000)
• seed: Random seed (-1 for random)
• enable_pbr: Enable PBR textures

Meshy - Text to 3D (Refine)

Refine a preview model with textures.

Parameters:

• preview_task_id: Task ID from the preview stage
• texture_prompt: Additional prompt for texturing
• enable_pbr: Enable PBR textures

Meshy - Image to 3D

Generate a 3D model from an input image.

Parameters:

• image: Input image
• ai_model: Model version
• topology: Mesh topology
• target_polycount: Target polygon count
• enable_pbr: Enable PBR textures
• should_remesh: Enable remeshing
• should_texture: Generate textures for the model

Output

Generated models are saved to ComfyUI/output/YYYY-MM-DD_HH-MM-SS/ with:

• 3D model files (GLB, FBX, OBJ, USDZ)
• Texture maps (base color, normal, metallic, roughness)
• Thumbnail image

Outputs from Nodes

| Output | Description | |--------|-------------| | thumbnail | Preview thumbnail of the generated model | | base_color | Base color texture map | | normal | Normal map | | metallic_roughness | Metallic/Roughness map | | model_path | File path to the downloaded 3D model |

API Pricing

Refer to Meshy Pricing for current API costs:

• Text to 3D Preview: ~20 credits (Meshy-6) / ~5 credits (others)
• Text to 3D Refine: Additional credits
• Image to 3D: Varies by model

Requirements

• ComfyUI
• Python 3.8+
• Valid Meshy API key

Dependencies

• numpy
• Pillow
• requests
• torch
• pyhocon

License

GPL-3.0 License

Links

• Meshy AI
• Meshy API Documentation
• ComfyUI

Troubleshooting

API Key Issues

• Ensure your API key is valid and has sufficient credits
• Check that the key is entered correctly without extra spaces

Generation Failures

• Check your internet connection
• Verify your prompt doesn't contain restricted content
• Ensure you have sufficient API credits

Model Not Downloading

• Check the output directory permissions
• Ensure sufficient disk space