Text-to-3D and image-to-3D
Two ways in: a prompt — "chrome skull with neon veins" — or a reference image from your set, a logo render, a photo. Combine both to steer the reconstruction with words.
AI · 3D Models
A mesh from a sentence — or a photo.
Describe an object and get a spinning 3D model in your library, or upload a reference image and the generator rebuilds it as geometry. Output is a standard GLB that drives the 3D Model source like any imported mesh.
Two quality tiers
Standard is the fast tier — a clean mesh in roughly 30–45 seconds, right for silhouette work, luma-keyed spins, and blocking out a look mid-set. PBR takes about 60–90 seconds and arrives fully textured with physically-based materials, so it holds up close on a big wall under the engine's lighting.
A regular model, the moment it lands
The result is a GLB in your media library — no special format, no AI inspector. Drop it in a clip, bind rotation and scale to the audio engine, light it, and re-skin it any time with an AI-generated PBR texture bundle. Long renders report into the Tasks panel so the deck never waits.
Tokens: ~20–40 per model depending on tier (standard vs PBR). Failed generations refund in full; cancel a queued job for a full refund.