Teams using AI in design reviews often need a fast way to move from a flat reference to a concrete visual asset. A useful workflow is to keep the input simple, create a first 3D draft, then refine the mesh or scene only when the concept is worth deeper work.
I have been collecting notes around Trellis 2, a Trellis AI style workflow for 2D to 3D, Photo to 3D, and prompt based 3D concept generation.
The useful pattern is:
- Start with one clean product photo, sketch, or prompt.
- Generate a quick 3D concept.
- Review scale, silhouette, and material direction.
- Only then export or rebuild the asset for production quality.
This keeps AI generation in the ideation lane while still giving product, design, and engineering teams a more spatial artifact to discuss.
Top comments (0)