← AI 3D workflows

GLB asset pipeline for WebGL and casual games

Atlas turns a short prompt into a lightweight, textured 3D asset with three ready-made LOD tiers — built for fast-loading WebGL and casual games, exported as GLB.

Try it on Atlas

Four stages, prompt to multi-LOD mesh

01Atlas canvas input node with the text prompt "running fat fluffy cat character"

Prompt in

Describe the subject — style directives (low-poly, hand-painted, claymation) get appended automatically to keep the 3D reconstruction clean.

02A stylized low-poly orange cat character concept render on a gray background, wearing a red bandana

Concept render

A clean 2D concept generates from the combined prompt, framed and lit for reconstruction.

03The cat character as a finished textured mesh, its normal map, and its wireframe, shown side by side

3D reconstruction

The concept reconstructs into a textured mesh with normal maps baked in, then decimates down to a base target around 5,000 triangles.

04

LOD generation

Origin, scale, and orientation normalize automatically, and two lighter tiers generate in parallel — around 3,000 and 1,000 triangles — for objects at a distance.

Where the time goes

StageManualAtlas
Text-to-image concept4–8 hours5–15 seconds
Image-to-3D reconstruction3–4 hours30–60 seconds
Decimation, transform & LOD tiers4–6 hours15–45 seconds
Total, prompt to multi-LOD mesh2–3 days1–2.5 minutes

Common questions

What does this workflow output?
A textured 3D model with UV mapping at three polygon budgets — roughly 5,000 / 3,000 / 1,000 triangles — exported as GLB with baked textures, ready for real-time use.
How long does it take?
1 to 2.5 minutes from a fresh prompt to all three LOD tiers. Re-running a later step with unchanged upstream inputs loads instantly from cache.
Can I control the art style?
Yes — change the subject in the prompt, layer on style directives (hand-painted fantasy, claymation, chibi anime, sci-fi hard-surface), or adjust the 3D backend and mesh settings for a different budget.
Which engines are supported?
GLB is a native fit for WebGL frameworks — Three.js, Babylon.js, PlayCanvas — and drops just as easily into Unity, UE5, Godot, and Blender, Maya or 3ds Max.

The stages in depth

The AI 3D asset pipeline, explained →AI retopology for game assets →

Related workflows

2D concept art to 3D game asset →Character image to rigged, animated 3D character →Game design document to concept art and poster →

Try this workflow in Atlas

Free to start. No credit card required.

Try it on Atlas