microsoft/TRELLIS ? reverse-engineered prompt

Build me a local app for turning a text prompt or a few reference images into a high quality 3D object. I want to be able to type something like “a ceramic cat statue” or upload images, then generate a usable 3D asset and preview it in the browser. It should also let me export the result in a few common forms like a mesh and other 3D formats if possible, and give me an easy way to make a variant or a small local edit of the same object.

Please include a simple web demo, a few example scripts I can run from the command line, and a clear setup so I can get it working on a Linux machine with an NVIDIA GPU. If there are current model or install details that matter, look them up online and use the latest recommended approach.