Skip to content

Flex3D

Build 3D assets from one image or a prompt

Visit official site

What Flex3D is

The method is described as a two-stage pipeline that produces high-quality 3D assets from one image or a text prompt, with support from Meta.

What you can do with it

  • Reconstruct a detailed 3D model from one image
  • Filter input views to refine the asset
  • Guide generation with a text prompt

Application scenarios

  • Generate a game environment quickly
  • Build a 3D scene from concept art
  • Create a detailed character from a sketch

Pros & cons

✓ What we like

  • Text prompts can guide the generation
  • Results are shown in an interactive viewer

! What to watch out for

  • The process runs in two separate stages
  • Input starts from one image or a text prompt

FAQ

What are the two stages?

Flexible reconstruction and input view filtering.

Who supported development?

The material states development was supported by Meta.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review