Project / 03 · MetaHuman development

VIVI.

A custom real-time fashion model and virtual performer built to explore facial performance, clothing presentation and the increasingly blurry border between character and presence.

  • MetaHuman
  • Unreal Engine
  • Live Link
  • Real-time
Close three-quarter portrait render of Vivi
PORTRAIT STUDY / 2026REAL-TIME RENDER

THE BRIEF

A believable digital model
with enough personality to talk back.

Vivi began as a MetaHuman fashion model for presenting clothing assets, then rapidly became something more interesting: a test bed for real-time facial performance and a possible visual identity for an AI companion.

The design process focused on breaking away from the untouched MetaHuman default while preserving its high-fidelity deformation system. Proportions, facial structure, skin, eyes, hair, makeup and body shape were iterated while testing expressions through Live Link.

Development notes

Built to move,
not merely pose.

01

Identity through iteration

Facial proportions, head scale, eye treatment, hair and makeup were repeatedly adjusted to give Vivi a distinct identity while working within MetaHuman’s available controls.

02

Real-time expression

Live Link facial capture turned static design decisions into performance problems. Smiles, concern, surprise and quieter neutral moments became part of evaluating whether the character felt alive.

03

Fashion-ready foundation

The body was shaped as a tall runway-style base for developing and presenting reusable clothing assets, with future work planned around styling, footwear and final presentation.

PROJECT STATUS / ONGOING

Next: clothing,
performance and presence.

RETURN TO SELECTED WORK ↗