VTuber Workflow
VTuber & Live2D Layer Separation: Turn One Drawing into Rig-Ready Layers
Live2D rigs and parallax animation both need character art split into separate layers. This guide covers what to split, how to do it with AI, and what to patch before rigging.
Why rigging starts with layers
A Live2D model deforms hair, eyes, mouth, body, and props independently. Given one flat PNG, the rigger can only slice along visible contours — and occluded regions (the shoulder behind the hair) are simply missing, which shows as soon as the model moves.
Proper practice starts with a layered PSD: one layer per movable part, with occluded areas drawn or filled in. AI separation turns that multi-hour manual prep into a single upload.
The standard split
Hair (front, back, side pieces)
Hair moves the most in a Live2D rig. The AI separates visible hair masses into their own layers; occluded roots still need hand-drawn fill-in afterwards.
Face (eyes, mouth, brows, nose)
Blinking and talking need eyes and mouth independent of the face base. After splitting, fill in the sclera behind lids and the lower lash line for closed-eye states.
Body and limbs
Torso, arms, and hands need their own layers for leaning and waving. Regions hidden by hair or props must be painted back in.
Props and background
Microphones, glasses, hats, and accessories on separate layers can move or be removed independently. A preserved background doubles as a streaming card or thumbnail plate.
Before the files go to the rigger
Patch the occluded regions
AI splits only visible pixels. Every region another layer covers — the shoulder under hair, the torso behind an arm — needs inpainting or hand-drawn fill.
Name layers and keep the order
Live2D Cubism imports PSD layer names directly. Clear names (hair_front, eye_L, arm_R) save the rigger real cleanup time.
Deliver a PSD, not loose PNGs
Riggers need layers aligned on one canvas. Loose PNGs in a ZIP have to be manually positioned; a PSD keeps the composition intact and opens ready to rig.
Start splitting
Split your character art into rig-ready layers
Upload character art, let the AI separate hair, face, body, prop, and background layers, then export the PSD for Live2D.
Open the Anime Layer Splitter