Which engine actually fits?
Auto Prep will ask you this and pick a sensible default — this is the same logic in plain language, for when you want to decide yourself.
A logo, crest, or clean line art
Use Shape. It carves from the silhouette, so sharp flat edges stay sharp instead of being read as flat gray tone.
A logo or crest with shading or texture on it
Use Blend. It layers a grayscale-luminance read on top of Shape's silhouette dome, so shading, engraved-style detail, or gradient fill survives instead of being flattened to a single depth. Tune Detail Strength and the Luminance Curve to taste — no credits used.
A real photograph, a portrait, or just an idea
There's no free local engine for a real photograph — Depth2Beam detects one on upload and blocks Shape and Blend for that file rather than forcing it through math built for flat artwork. Use the Relief Generator instead: describe the subject and get a sculpted medallion-style depth map with no image required, then switch on the AI Depth refinement pass (or AI Depth + Detail for finer surface texture) afterward, and the cloud model understands facial and anatomical structure in a way flat luminance can't. There's no manual AI Depth engine to point an uploaded photo at directly — this generate-then-refine path is the way there.
Let Auto Prep decide for you
Tell the console whether your image is a Logo/Crest or Mixed, and it steers you toward the right local engine automatically — a real photo or portrait upload is flagged and blocked instead of forced through the wrong math.
Launch Depth2Beam →