ComfyUI Extension: Tripo for ComfyUI

Authored by VAST-AI-Research

Created

Updated

200 stars

Custom nodes for using a/Tripo in ComfyUI to create 3D from text and image prompts.

README

ComfyUI-Tripo

This extension integrates Tripo into ComfyUI, allowing users to generate 3D models from text prompts or images directly within the ComfyUI interface.

ChangeLog

  • 20241111: adapt for new api
  • 20241014: support convert
  • 20240913: support model_version v2.0-20240919

Features

  • Generate 3D models from text prompts
  • Generate 3D models from images
  • Animate 3d models
  • Convert format and retopologize
  • View generated 3D models directly in ComfyUI using the TripoGLB Viewer node

Installation

[method1] From Source

  • Clone or download this repository into your ComfyUI/custom_nodes/ directory.
  • Install the required dependencies by running pip install -r requirements.txt.

[method2] From ComfyUI-Manager

If you have ComfyUI-Manager, you can simply search "Tripo for ComfyUI" from Custom Nodes Manager and install these custom nodes

[method3] From Comfy Register using comfy-cli

If you have a comfy-cli, you can simply execute comfy node registry-install comfyui-tripo in command line.

Usage

How to get a key

  • Generate an api key from Tripo
  • Set your key by:
    • [Method1] Set your Tripo API key as an environment variable named TRIPO_API_KEY in your env variables.
      • Windows
        set TRIPO_API_KEY=tsk_XXXXXXX
        python.exe main.py [--cpu]
        
      • Linux/Mac
        TRIPO_API_KEY=tsk_XXXXXXX python main.py [--cpu]
        
    • [Method2] Set your Tripo API key in node input field.
    • [Method3] Set your Tripo API key in config.json.

Usually it will take 10~15s to generate a draft model.

How to use a workflow

Load the png sceenshot in comyfui by dragging or loading manually.

Workflows

Text to Mesh

img

This node allows you to generate a 3D model from a text prompt.

Image to Mesh

img This node allows you to generate a 3D model from an input image.

Multiview Images to Mesh

img This node allows you to generate a 3D model from three multiview input images.

Texture a generated model

img This node allows you to generate texture and pbr for a generated 3D model.

Refine a draft Mesh

img This node allows you to refine a 3D model from a draft model.

Animation

img This node allows you to generate a 3D model with skeleton and animation.

Download Model

Models will be automatically downloaded after generation in ComfyUI\output folder.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Credit

Thansk for initial repo from Tripo-API-ZHO