Skip to content

Make-A-Character

Build lifelike 3D characters from text descriptions

Visit official site

What Make-A-Character is

The framework is described as a user-friendly way to create lifelike 3D avatars from text descriptions, with characters that are controllable, realistic and fully realised within two minutes.

What you can do with it

  • Create a 3D avatar from a text description
  • Build a controllable and realistic 3D character
  • Make a fully realised character within two minutes
  • Integrate the character into existing CG pipelines for dynamic expression

How the framework works

  • Large language and vision models interpret text intent
  • The same models generate intermediate images
  • Human-oriented visual perception modules follow
  • 3D generation modules produce the final character

Character qualities

  • Output is described as controllable
  • Output is described as highly realistic
  • The character is complete rather than partial
  • The character is animatable and industry compatible

Pros & cons

✓ What we like

  • Characters are described as controllable and fully realised
  • Dynamic expression is supported through CG pipeline integration

! What to watch out for

  • The workflow starts from a text description
  • The pipeline relies on large language and vision models

FAQ

How long does a character take?

The material states a controllable and lifelike 3D character can be made within two minutes.

Can the result be animated?

The feature list names animatable output, and the material mentions dynamic expression through CG pipeline integration.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review