Adaptive Auto Skinning

Automatic skinning for your characters: the body and every clothing mesh bound to clean weights in about a second, fingers included, on your GPU.

Clean weights on a full character usually cost hours of weight painting, or a fight with Automatic Weights. Adaptive Auto Skinning binds the rig, body plus clothing, in one click, with every vertex weighted. It is volumetric skinning reinvented for fingers and tight zones: the grid stays coarse through the body and refines exactly where two surfaces almost touch.

A real bind from Adaptive Auto Skinning, no manual paint. Each color is one bone.

The whole body in one pass
Point it at the armature. It finds every mesh the rig deforms, solves them together, and writes standard vertex groups back.

Sub-second binds on your GPU
The solver runs on your graphics card, on macOS, Windows, and Linux alike. Across the 202-character benchmark set, the average full-rig bind took 0.57 seconds, measured on a MacBook Air M4.
Scored against artist weights
On the 202-character benchmark, its weights pick the same dominant bone as the artist's hand-painted reference on 80.9% of vertices, ahead of every automatic method tested. The full benchmark, every method on 202 characters with the per-character tables, is on the blog.
Any rig, any bone names
It reads the skeleton's geometry, not its names. A Mixamo rig, an Auto-Rig Pro skeleton, and a hand-built armature all bind the same way, with no naming convention to follow. Face rig still requires manual work, an automatic feature is in progress.