How to Make a Vtuber Model in Procreate
StyleFrame Team7 min read
To make a VTuber model in Procreate, draw the character in separate parts, export each part as a layered PSD, merge the PSDs into one file on a desktop in Photoshop or Clip Studio Paint, and then rig that file in software like Live2D Cubism. Procreate's layer limit depends on your iPad, canvas size, and resolution, so many artists split the character into two to five canvases (head, body, hair, and so on) to get enough layers. Procreate handles the drawing only. The movement comes from the rigging stage.
What a VTuber Model Actually Requires
A 2D VTuber model is not a single illustration. It is a character cut into many pieces: eyelids, irises, eyebrows, mouth shapes, front hair, side hair, back hair, arms, and any accessories that need to move. A rigging tool deforms and animates those pieces, and a face tracker drives them.
If you come from motion design, think of it as a character prepared for a puppet rig in After Effects. Every part that moves independently needs its own layer, and anything hidden behind another part has to be drawn so it doesn't reveal a gap when the top part moves.
That is why the process is organized around layer management rather than just drawing. Procreate is a good illustration tool, but it caps layers per canvas, so the workflow has to account for that.
Step 1: Plan the Design and Choose a Scope
Sketch in a notebook or a loose Procreate canvas before committing to a final design. Gather references for palette and costume, and decide how complex the model will be.
If this is your first model, start with a chibi or half-body design. Fewer parts means fewer layers, less hidden-area painting, and a simpler rig. You can build up to a full-body model after you've completed one end-to-end.
Decide early where the model will be used. A streaming model that mostly shows head and shoulders needs far less art below the chest than a full-body model for an animated video.
Step 2: Set Up the Procreate Canvas
Canvas size controls how many layers you get, so set it deliberately.
- Width: Make it larger than 3,000 px so the detail holds up when the model is scaled in the rig and on stream.
- Height: Match your character's proportions. Around 4,000–5,000 px is a common range for a full-body figure.
- Layer count: Check the layer count Procreate offers for your dimensions. Aim for more than 10 layers per canvas at minimum, and expect to trade resolution for headroom.
Next, turn on symmetry. Open Actions > Canvas > Drawing Guide > Edit Drawing Guide > Symmetry and finish. Enable Assisted Drawing on the layer you're working on. You can hide the guide lines afterward, and the layer stays symmetrical as long as it still says "Assisted." New layers don't inherit this setting, so many artists map Assisted Drawing to a gesture or tap shortcut under Gesture Controls to save time.
Symmetry helps with eyes, hair mirrors, and clothing. Turn it off for asymmetric parts like bangs and accessories.
Step 3: Sketch the Whole Character on One Canvas
Draw the full character first, with all the detail and flat colors you need, before you cut anything apart. It is much harder to adjust proportions, palette, or costume details once the model is divided across files.
Treat this master sketch as the source of truth. When it's done, merge its layers, or duplicate the file first if you want to keep the originals.
One practical tip: Live2D can slightly desaturate PSDs exported from Procreate (one creator estimates about 10%). Raise the saturation on the master sketch before you split it, so the colors come out right in the rigging software. Check this in your own pipeline, since results depend on color profile and export settings.
Step 4: Split the Model Across Canvases
For each section, duplicate the master file and crop it to that area: head, torso, legs, and so on. Cropping removes empty space, which frees up room for more layers. Turn on snapping to center a crop cleanly.
Inside each file, keep two copies of the reference sketch:
- one at low opacity for tracing,
- one at full opacity, hidden or locked, for color picking.
Give each section the same scale and consistent alignment so the pieces assemble cleanly later. Note the pixel offset of each crop, or use a consistent grid, so you can place each PSD at the right position on the final canvas.
Step 5: Draw and Separate the Parts
Work through each canvas, drawing parts on separate layers and finishing the hidden areas that the rig will reveal.
For the head, you typically need:
- Eyelashes, upper and lower lids, whites, irises, and highlights as separate layers
- Eyebrows
- Mouth shapes (closed, open, and any teeth or tongue)
- Face base, nose, and blush
- Front hair, side hair, and back hair split into movable chunks
- Ears and any head accessories
For the body, separate the neck, torso, each arm segment, hands, and anything that should sway, such as ribbons or sleeves.
Where one part overlaps another, draw the covered portion. For example, extend the hair beyond the face edge, complete the neck beneath the chin, and paint the full eyeball behind the eyelid. Leaving these areas blank is the most common reason a rig looks broken once it moves.
Name layers clearly as you go. Hundreds of layers named "Layer 47" will slow you down in the rigging stage.
Step 6: Export PSDs and Merge Them on Desktop
Export each canvas as a PSD from the Share menu. Then open them on a desktop in Photoshop or Clip Studio Paint and combine them into one file by copying layers into a single master document at the canvas size and positions you planned.
Before you move on, check:
- Layer names and order survived the export.
- Blend modes look the same as they did in Procreate. Complex modes can render differently.
- Color matches the original. Fix any shift now rather than after rigging.
- Edges are clean, with no stray fringes on parts that need to deform.
Keep the final merged PSD as your master and save versioned copies. Keeping layer names consistent will make it easier to revise the art later.
Step 7: Rig the Model
The usual route is Live2D Cubism. This is where the model learns to blink, turn its head, and move its mouth.
If your goal is pre-rendered animation rather than live streaming, you may be able to take the layered PSD into other animation software and rig it there instead.
Whichever you choose, test blinking, head turns, and mouth shapes early, then go back to Procreate and fix any areas where the art reveals gaps.
Where Styleframe Fits in a VTuber Workflow
Styleframe doesn't replace the layered drawing described above. A rig needs clean, separated parts, and that remains hand-built work in Procreate and your PSD editor.
The aim is to keep your existing pipeline and use any additional tools only where they speed up a specific step.
Common Mistakes to Avoid
- Making the canvas too small. Low resolution limits detail and scaling, but too large a canvas leaves almost no layers. Test the layer count before you start.
- Cutting the model apart too early. Finish the master sketch and color decisions first.
- Skipping hidden areas. Gaps show up the moment a part moves.
- Inconsistent crops. If sections don't align, merging PSDs becomes tedious.
- Poor layer naming. It costs time at every later stage.
- Ignoring color shifts. Check the export in your rigging software and compensate before final art.
- Taking on too much. A finished chibi beats an abandoned full-body model.
Final Thoughts
Procreate can produce a complete VTuber model, but only if you plan for its layer limits. Set a sensible canvas size, draw the whole character once, split it across several files, separate every moving part, merge the PSDs on desktop, and rig in Live2D. Done in that order, the workflow is manageable, and each stage is easy to check before you move on to the next.