AI changed what a "3D model" feels like to make. You no longer need to learn Fusion 360 or Blender to design a custom bracket, a figurine of your dog, or a terrain map for tabletop gaming. But "AI 3D" isn't one thing — it's three distinct workflows (text-to-3D, image-to-3D, text-to-CAD) that each shine in very different situations. This page is the map.
Organic shape from a description → text-to-3D. Recreate a real object you can photograph → image-to-3D. Engineered part with exact dimensions → text-to-CAD. Subject fits a known category (car, house, pet, face, city) → use the specialized generator for that category — it almost always beats the general model.
You type a description ("a chunky low-poly owl, sitting, no base"). A generative model produces a textured 3D mesh in seconds — an STL, OBJ, or GLB you can drop into a slicer. The underlying tech is usually a diffusion model that paints many views of the subject, then reconstructs a mesh from those views.
You upload one or more photos of an object. The model infers its 3D shape and produces a mesh that approximates it. Under the hood it's the same generative pipeline as text-to-3D, with the photos acting as a strong visual prompt.
See Image-to-3D best practices and Photographing objects for image-to-3D for the photo workflow that actually produces clean results.
You describe a part ("a wall-mount bracket for a 50 mm router with M4 screw holes on 60 mm centres"). The AI generates parametric code (in our case, OpenSCAD) that builds the shape from variables. You can edit any variable — or ask the agent to — and the model rebuilds. Want the wall 2 mm thicker? Change one number.
Walk down the list until you hit a "yes":
The most powerful workflow today is text-to-3D for the artistic shape , then text-to-CAD for the functional base or mounting interface . AI sculpts the figurine; the CAD Agent designs a perfect-fit display stand. Export both as STL and assemble them in your slicer.
Generative 3D models burn real GPU time, which is why every provider (PrintPal included) meters them with credits or per-generation pricing. A handful of practical implications: