Sculpt Layers Explained: The Two Kinds You'll Meet

Two people can say "sculpt layers" and mean completely different features. One means a stack of recorded deformations you can dial up and down like an opacity slider. The other means separate mesh parts you can move, hide, and merge. Knowing which one a tool gives you saves a lot of confused searching.

What are sculpt layers?

Sculpt layers are a way to keep parts of a sculpt separate instead of baking every stroke into one mesh. The term covers two distinct features: deformation layers, which record strokes as an adjustable offset from the underlying surface, and part layers, which hold separate mesh objects in one scene. ZBrush and Nomad Sculpt have the first kind; many tools, including Re:Form, have the second.

The confusion is not anyone's fault. Both features live in a panel called "Layers," both let you toggle things on and off, and both get called non-destructive. But they solve different problems, and swapping one for the other in your head leads to disappointment when you go looking for a slider that isn't there.

The two kinds, side by side

The mental model that helps: a deformation layer is a transparency sheet laid over the same drawing. A part layer is a different drawing on the table.

Deformation layers can only exist where the app has a fixed surface to measure against. That's why they come with rules about subdivision levels — the offset has to be stored relative to something. Part layers have no such constraint, because each one is just its own geometry.

How do sculpt layers work in ZBrush?

ZBrush 3D Layers store geometry, polypaint, and masking, and they only accept changes while Record mode is on — you click the REC icon to start, and click it again to stop. Per Maxon's documentation as of August 2026, new layers must be created at the highest subdivision level, and you must return to that top level before you can switch recording off.

That last rule catches people out constantly. You start recording, subdivide down to work on the big forms, and then find the REC toggle won't release. The fix is not a bug report; it's stepping back up to the top level first.

What you get in exchange is the thing deformation layers are actually for: an intensity slider. Sculpt a set of wrinkles onto their own layer and you can pull them to 50% when they read as too heavy, or push past 100% when they read as too soft — without re-sculpting anything. The same mechanism turns layers into crude blendshapes: put a smile on one layer and a frown on another, and the sliders mix them.

Nomad Sculpt works on the same principle. Its manual describes layers as storing edits non-destructively, with per-layer opacity, blend modes, duplication, and reordering, and it explicitly frames sculpt layers as working like blendshapes or morph targets for things like facial expressions. Nomad also runs color, roughness, metalness, and opacity channels through the same layer stack, with a per-channel factor multiplied against the layer's overall strength.

Does Blender have sculpt layers?

Blender has no native sculpt layer stack as of August 2026 — not in 5.2 LTS, and the feature is still an open design discussion rather than shipped code. Blender's non-destructive sculpting instead runs through the Multiresolution modifier and Shape Keys, plus paid third-party add-ons that reconstruct a layer stack on top of those primitives.

This is worth understanding rather than complaining about, because the two built-in options do overlapping jobs badly if you pick the wrong one.

Multiresolution stores your sculpted detail per subdivision level, so you can drop to a lower level, adjust the large forms, and step back up with the fine detail intact. That is genuinely non-destructive, but along one axis only: resolution. It gives you no way to fade a set of wrinkles to half strength.

Shape Keys are closer to what people want from deformation layers. Each key stores a deformed version of the mesh with a 0–1 slider, so you can toggle a sculpting pass on and off, or blend two passes. The catch, well documented in Blender's own tracker and community threads, is that Shape Keys and the Multiresolution modifier do not coexist: there's no multires shape key support, only shape keys on the base mesh, so you have to apply multires first. You get detail or blending, not both at once.

That gap is exactly why paid add-ons for sculpt layers exist and sell — the listings for them state plainly that no official Blender feature does the same job. If deformation layers are load-bearing in your workflow, budget for an add-on or plan around Shape Keys on a base mesh. We wrote about where Blender wins and loses generally in our Blender comparison; for sculpting breadth, it still wins most rounds.

How Re:Form's layers work

Disclosure first: Re:Form is our tool, so weigh this section accordingly. Re:Form's Layers panel is a part layer system, not a deformation layer system. Each layer is its own mesh you can move and sculpt independently, and there is no intensity slider — nothing here fades a stroke to 40%.

What the panel actually gives you, on the Free plan as much as Pro:

• Add — creates a new layer starting from a primitive you pick, so a fresh sphere lands in the scene ready to become an eyeball or a horn.
• Duplicate — an exact, independent copy of the active layer.
• Mirror — flips the active layer across the X axis in place. Duplicate first if you want a mirrored copy, which is the two-click way to place a second eye.
• Move ↑ / Move ↓ — reorders the row in the list. This is list order only; it changes nothing about the sculpt. There's no stacking or blending happening between rows.
• Merge down — bakes the active layer into the one directly below it, keeping the lower layer's name. The two become one mesh.
• Merge all — flattens every visible layer into a single mesh. Do this before remeshing the whole model as one solid.

So, plainly: if you came looking for ZBrush-style recorded deformation with an intensity dial, Re:Form does not have that, and this post is not going to pretend otherwise. If you came looking for a way to build a model out of separate movable parts and flatten them when you're done, that is exactly what's there, and it's not gated behind a plan.

Which kind do you actually need?

Most beginners need part layers and think they need deformation layers. Building a head out of a skull, two eyeballs, and separate ears is a part-layer job, and it's the one you'll hit in your first week. Deformation layers earn their keep later, on corrective passes and expression work — real production needs, but not first-project needs.

Here's the practical test. Ask what you want to do to the work after you've done it:

• "I want to move it, hide it, or delete it." Part layers. A separate eyeball you can nudge into the socket after the fact beats one carved into the skull, because the carved version can't be un-carved.
• "I want less of it." Deformation layers. Skin pores at 60%, a brow furrow dialed back until it stops reading as anger. Nothing in a part-layer system does this.
• "I want two versions and I'll pick later." Either, via duplication. Copy the layer, experiment on the copy, delete the loser.

The third case is the one worth internalizing, because duplication covers a surprising amount of what people reach for layers to do. It costs you memory instead of elegance, but it works in every tool ever made.

Working non-destructively without deformation layers

If your tool doesn't have an intensity slider, you're not stuck — you just spend a little more discipline instead. Four habits do most of the work.

Duplicate before anything risky. Trying a different nose, an experimental crest, a jawline you're unsure about? Duplicate the layer first, sculpt on the copy, keep whichever version survives the next morning. This is the poor man's layer stack and it's genuinely fine.

Mask instead of committing. Masking protects geometry from strokes, which means you can push hard on one region without touching its neighbor. It doesn't record anything, but it prevents the damage you'd otherwise be trying to undo. Worth pairing with an understanding of what each brush actually does — we covered that in the brush guide.

Save real files, not just autosaves. Browser autosave is a safety net for closed tabs, not a version history. Writing named files at milestones — head_v1, head_v2_newjaw — gives you the one thing no layer system does: the ability to go back a week. In Re:Form that's a local .clay file, which snapshots the whole session including every layer.

Merge late. Flattening is a one-way door. Keep parts separate as long as they're useful and merge only when something downstream demands a single mesh — remeshing the model as one solid, or exporting a watertight body for a printer. There's no prize for a tidy layer list.

That last point is where part layers and printing collide, and it's worth planning for. A model made of five floating parts looks fine on screen and confuses a slicer. Merge all, then check the result, then export.

Where to start

If you want to feel the difference rather than read about it, the fastest route is to build something with obviously separate pieces — a head with distinct eyeballs is the classic, and we walk through it in the head-sculpting workflow. You'll understand part layers in about ten minutes of moving an eyeball around, which is faster than any explanation.

You can try it in a browser tab on the Free plan: every brush, sculpt layers, masking, symmetry, and remeshing, up to roughly 12.5 million triangles, with OBJ import and export and no watermark. It needs a free sign-in and a desktop browser with WebGPU — current Chrome, Edge, and Firefox, plus Safari 26 on macOS Tahoe. Re:Form is v0.6.7 Alpha, so expect the occasional rough edge; more detail sits in the FAQ.

FAQ

Can you convert part layers into deformation layers?

No — they store fundamentally different data, so there's no conversion path in any tool. A deformation layer holds an offset measured against a base surface; a part layer holds independent geometry with no base to measure against. Moving between apps, you export meshes, which means part layers survive as separate objects and deformation layers get flattened into whatever the intensity sliders were set to at export time.

Does Re:Form have ZBrush-style layers with an intensity slider?

No. Re:Form's Layers panel holds separate mesh parts — Add, Duplicate, Mirror, Delete, reorder, Merge down, Merge all — with no per-layer intensity or blending. Full disclosure that Re:Form is our tool: if recorded deformation with an adjustable strength dial is central to your work, ZBrush or Nomad Sculpt currently do that and Re:Form does not. The ZBrush comparison covers the rest of that gap.

Do sculpt layers survive an OBJ export?

Partly, and it depends which kind. Part layers export as separate objects or groups, so a head, two eyes, and a horn arrive in Blender as distinguishable pieces. Deformation layers do not survive at all — the exporter writes the final surface with every layer at its current intensity, and the stack itself is gone. If you need the stack, keep the app's native project file alongside the export.