Features
Everything the platform does, one page each — what you get, how it runs, and what it cannot do. 18 features.
Models
The engines underneath — what each one is built for.
Zetta N1Our original neural model. Text, a single image, or up to seven reference views become a mesh — plus low-poly, sketch and untextured modes that no other model here offers.Zetta N2A second neural engine with a deliberately narrow surface: exactly one image, textured output. Fewer knobs, a different reconstruction, and available on paid plans.Zetta P1Rather than reconstructing a shape from pixels, P1 reasons about structure and generates geometry as a set of operations. Built on top of Claude, and suited to anything with rules.
Generation
Getting from an idea, a photo or a set of views to a mesh.
Text to 3DA sentence becomes a reference image, and the image becomes geometry. The fastest way from an idea to something you can rotate.Image to 3DThe most predictable route into 3D. You choose the silhouette instead of asking a model to invent one.Multi-view to 3DUp to seven reference views alongside the primary image, so the far side of your model is observed rather than inferred.
Refinement
Turning a generated mesh into something you can ship.
RetopologyReconstruction produces dense, irregular triangles. Retopology keeps the silhouette and rebuilds the mesh at a density you choose, in quads or triangles.AI SegmentationSplit a model into separate named pieces — a chair into seat, back and legs — each downloadable on its own or together as one file that arrives already divided.UV UnwrappingGenerated UVs are a by-product — fragmented, unevenly scaled, seamed wherever the algorithm found convenient. Unwrapping gives you a real layout, and shows you what it did.AI TexturingChange a surface without touching the shape — driven by a description or a reference image, previewed as a picture before a single credit goes on the mesh.PBR ChannelsGenerate normal, roughness, metalness, emissive and ambient occlusion from a model’s base colour — each at the source texture’s own dimensions, so it lines up pixel for pixel.AI RiggingBones and skin weights for humanoid characters — with a pre-flight check that looks at your model and tells you if it cannot be rigged, before anything is spent.Asset TaggingTags generated from what the model looks like, from a controlled vocabulary — so “vehicle” always means vehicle and a filter finds every one of them.
Platform
Bringing your own work in, and driving all of it from code.
Bring Your Own ModelUpload a GLB and it becomes an ordinary library model — retopology, unwrapping, texturing, rigging, segmentation, tagging and conversion all run on it. Nothing was generated, so nothing is charged for the upload.Format ConversionGLB, OBJ, FBX and STL, converted from any model in your library — without a round trip through a DCC application to change a container.Background RemovalA cluttered background is the most common cause of a bad reconstruction. Cutting the subject out before it is submitted costs almost nothing and changes the result more than any other setting.Image StudioGenerate, edit and compose images in the same product that turns them into meshes — so a reference you are happy with is one click from a generation instead of one export away.Library & SharingOne library across the studio, the agent and My Generations — searched by id, title and tag at once, filtered by what made it, and shareable as a link anyone can open.