Diffusion · 3D · Three.js· in progress
Generative 3D Character
A pipeline that takes a short text prompt and produces a riggable 3D character. Diffusion-based generation feeding into retopology and skeleton handling, served in the browser through Three.js.
Three.jsWebGLDiffusionPython
The idea
Asset creation for games and simulations is still expensive. A pipeline that turns text into a usable, animatable model isn't sci-fi anymore, but the path from "diffusion output" to "asset a developer would actually ship" still requires real engineering. This project is me walking that path end to end.
Status
In progress. The generation stage works. Retopology and rigging are the hard parts. More to come.