Skip to content
← The showcaseComplete free read
Creator thread

Coordinate a 3D character’s face, gaze, and body per message

A character-performance experiment asks ten questions per message, then maps probabilistic emotion and body-region choices onto an existing rig.

Editorial workflow map
  1. Read the message
  2. Choose expression controls
  3. Compose the live reaction
View the author’s original post (opens in a new tab)
Our interpretation · not the author’s screenshot or a reproduced result

One request, ten bounded questions

The creator describes asking for an overall emotion first, followed by nine physical regions. Each question uses plain-language levels with probabilities rather than generated animation instructions.

Supporting documentation (opens in a new tab)

Application code owns the rig

A same-author continuation says application code maps the selected levels to rig controls and drops contradictory choices before composing the reaction.

Supporting documentation (opens in a new tab)

What remains unverified

The rig assets, mapping code, runtime behavior, and full media playback were not independently evaluated. The thread does not establish that Jev synthesized the character, animation assets, or speech.

Use the same checks on another demo.

Read the complete free evaluation guide →