mrforexample/comfyui-3d-pack

ComfyUI-3D-Pack is a plugin for ComfyUI, the popular node-based image generation interface. While ComfyUI is mainly known for generating 2D images, this pack adds a large collection of nodes that let you work with 3D objects. You can take a single photograph or image and turn it into a 3D mesh, apply textures to it, and export the result for use in games, animation, or other creative projects. The plugin brings in support for many different 3D generation models, each with its own strengths. Some are better at generating clean geometry from a photo, others can add color textures on top of an existing shape, and some can break a 3D object into separate named parts automatically. Models like Hunyuan3D, TripoSG, TRELLIS, InstantMesh, and PartCrafter are all available as nodes you can drag into a workflow and connect together, just like any other ComfyUI node. Installation is handled through ComfyUI-Manager, which is the standard way to add plugins to ComfyUI. There are also pre-built packages for Windows users on Python 3.12 with CUDA 12.4, so you do not need to compile anything yourself in most cases. Linux and Docker instructions are also provided. Some nodes do require Visual Studio Build Tools on Windows or gcc/g++ on Linux because they compile small pieces of code the first time they run. The repository includes a folder of example workflows in JSON format. These are ready-made pipelines you can load into ComfyUI to try out each model without building anything from scratch. The workflows cover single-image-to-3D, multi-view generation, texture baking, and scene decomposition into separate objects. This project is aimed at artists and developers who already use ComfyUI and want to add 3D generation to their pipelines. A basic familiarity with ComfyUI nodes is expected, there is no beginner tutorial included. The project is actively maintained, with new models added regularly.