The complete AI 3D pipeline.
Eight tools that work as one. Generate, texture, rig, optimize and publish 3D assets without bouncing between five different apps.
Brief it like an artist. Get it like a studio.
Prism-2 understands shape, material, era and mood. Ask for a "rusted 1950s diner sign, neon tubes, slightly bent" and you get exactly that, with real geometry behind it.
- Four candidates per prompt so you can pick the closest direction
- Style presets for realistic, stylized, low poly, voxel and toon
- Negative prompts and seeds for repeatable results
- Part regeneration to redo just the handle, the head or the wheels
One photo in. A full 360° model out.
Upload a product shot, a concept painting or a phone snap. Dimensa infers the hidden sides, separates materials and keeps proportions faithful to the source.
- Single or multi view input, up to 8 reference images
- Automatic background removal and lighting delighting
- Material separation for fabric, rubber, metal and glass
Your napkin doodle deserves a mesh.
Draw a rough silhouette in the browser or upload a scanned sketch. Add a short prompt for materials and Dimensa fills in the volume, detail and texture.
- Built in canvas with front and side view guides
- Respects your proportions instead of "improving" them
- Works with tablets and pressure sensitive pens
Any surface. Any material. Seconds.
Bring your own mesh or use one you generated. Describe the look and get a complete, physically based texture set that holds up under any lighting.
- Full PBR stack: albedo, normal, roughness, metallic, AO and emissive
- Up to 4K resolution with seamless UV seams
- Brush inpainting to fix or restyle a single area
- Advanced materials: clearcoat, sheen, transmission and iridescence
From statue to performer.
Dimensa detects humanoids, quadrupeds and props, then builds a skeleton with clean skin weights. Pick from 300+ motions or describe a new one in words.
- Engine ready skeletons compatible with Unity Humanoid and Unreal Mannequin
- Text to motion for custom idles, emotes and attacks
- Retarget any animation to any rigged character
Topology your tech artist will approve.
Set a triangle budget and Dimensa rebuilds the mesh with edge flow that follows the form. Deformation zones get extra loops so characters bend the way they should.
- Quad dominant or triangle output with a hard poly budget
- Automatic LOD chains from LOD0 to LOD4
- Normal map baking from the high poly source
FinalTopology
Not just objects. Whole worlds.
Describe a place and Scene Composer lays out dozens of consistent assets with matching style, scale and lighting. Rearrange anything, then export the full scene.
- Style lock keeps every asset in the same visual language
- Physics aware placement so props rest where they should
- Exports as USD or engine native prefabs
Let shoppers pick it up.
Embed an interactive 3D viewer with AR on any product page. Colorways, configurators and analytics included. Try it: this camera is live.
- One line embed plus native apps for Shopify and WooCommerce
- View in your space with USDZ and Scene Viewer
- Engagement analytics: rotations, AR launches, add to cart
Prism-2, a model trained to build.
Most 3D generators produce pretty point clouds. Prism-2 was trained end to end on licensed, artist made assets to output real meshes with UVs, materials and structure.
Native mesh output
Generates watertight geometry directly instead of converting from noisy volumes.
Licensed training data
Every training asset is licensed or commissioned, and artists are paid royalties.
Material aware
Predicts physically based materials per part rather than baking lighting into color.
Real world scale
Objects arrive at believable sizes in meters, with pivots on the ground plane.
Twelve export formats
Pick a target and Dimensa handles axis orientation, scale, texture packing and compression.
Native plugins
Generate and import without leaving your editor. Assets stay linked so updates flow through.