MULTI-MESH DECOMPOSITION

Split one 3D model into customizable parts

Uploading 25 separate 3D files for a customizable shoe or furniture piece forces the browser to run 25 network requests and lose shared assembly alignment. Aircada splits a single multi-mesh 3D upload into independent scene units that can be styled, hidden, or animated individually while sharing one cached network download and GPU buffer pool.

Split one 3D model into customizable parts

The tradeoff between one monolithic 3D file and dozens of part files

When building a multi-zone product customizer, teams often face an awkward choice. If they import a product as one monolithic 3D object, they cannot attach separate visibility rules, click triggers, or animations to the collar, sole, or hardware. Conversely, if they export 20 separate GLB files from Blender and upload them one by one, the browser fires 20 HTTP requests on startup and part origins easily drift out of alignment.

Aircada gives you both single-file simplicity and per-part control through sub-mesh unit decomposition and interactive viewport splitting. You upload one multi-part 3D file, and the engine automatically spawns an addressable scene unit for every layer and sub-mesh while backing all of them with a single reference-counted GPU media cache. Each extracted sub-mesh binds directly to option SKUs, conditional rules, and pricing so multi-part builds flow into Shopify or headless checkouts with accurate assembly lists for fulfillment.

IMPORT DECOMPOSITION & VIEWPORT SPLITTING

Turn one multi-part 3D file into independent scene units

When you import a multi-mesh model into Studio, Aircada wraps the parent file in a structured object set and extracts every named layer and sub-mesh as its own scene unit with exact local transforms and original material mappings intact. If a fused model still needs finer separation after upload, enter interactive Split Mesh mode in the viewport to click or drag-select specific sub-meshes and split them out without returning to CAD.

Automatically preserves nested layer hierarchies, relative pivots, USDZ orientation, and initial material assignments on import
Interactive viewport Split Mesh mode supports click and Shift/Ctrl multi-selection with live selected-part counters
Three non-destructive viewport actions: Split All, Split (extract selected and remove from source), and Duplicate (extract copy while keeping source)
Upholstered lounge chair separated into bent-walnut shell, tufted leather cushion, and brushed brass swivel base sub-mesh units
REFERENCE-COUNTED GPU CACHE

Fetch once over the network and share geometry in VRAM

Even if an uploaded assembly splits into 40 configurable sub-meshes, the browser downloads and parses the source binary only once. Aircada's live media cache deduplicates concurrent loads for the same asset URL, tracks every active unit holding a reference to that model or texture, and frees GPU buffers automatically once the last referencing unit unloads.

Single network fetch and GLTF parse pass; concurrent part requests attach to the in-flight load instead of downloading twice
Reference-counted holder tracking across both 3D geometry and texture cache pools prevents duplicate buffers in VRAM
Deterministic GPU buffer disposal fires automatically the instant an asset's active holder count reaches zero
Precision metallic hardware components arranged in a studio knolling layout representing a shared reference-counted GPU memory pool
WORKFLOW ADVANTAGE

One export for 3D artists, individual part ports for Studio builders

Keep your 3D asset pipeline clean without sacrificing granular configurator logic.

3D ARTIST WORKFLOW

Export the entire assembly as a single file

Artists model and align all cushions, frames, stitching, and hardware inside a single Blender or CAD scene and export one file. There is no need to batch-export dozens of individual part files or manually re-enter XYZ offsets in the browser.

STUDIO LOGIC WORKFLOW

Wire rules, materials, and click events to any sub-mesh

Inside Studio, every sub-mesh appears as a first-class unit. Merchandisers can hide a hood when a collar option is chosen, animate a single drawer front, split out a fused hardware trim directly in the viewport, or let shoppers click on a watch bezel to open its finish menu.

Verified mesh decomposition, viewport splitting, and GPU caching behaviors
Capability StageVerified Engine & Studio BehaviorOperational Benefit
Automatic Import DecompositionCreates scene units for every mesh and nested layer, preserves local position/rotation/scale, corrects USDZ up-axis orientation, and collapses redundant single-root wrappers.Zero manual re-alignment after importing complex assemblies from Blender or CAD.
Viewport Split Mesh ModeClick or Shift/Ctrl drag-select highlighted sub-meshes with live selection count; supports Split All, Split (cut), and Duplicate (copy).Separates late-discovered customization zones inside Studio without round-tripping to 3D artists.
Request DeduplicationConcurrent unit loads referencing the same GLTF or texture URL join the single in-flight fetch promise.Cuts startup HTTP overhead on multi-part products from dozens of requests down to one.
Reference-Counted DisposalTracks active holder count per cached asset URL and invokes GPU disposal automatically when active holders reach zero.Prevents VRAM leaks when shoppers browse across multiple configurable models on mobile.

Products built with multi-mesh decomposition

Ideal whenever a single product has multiple customizable zones, swap-in accessories, or moving parts.

Custom footwear & team uniforms

Split a sneaker or jersey into 8 to 15 colorable panels, laces, eyelets, and soles that all load from a single compressed GLB payload.

Upholstered furniture & seating

Separate wooden frames, seat cushions, piping welt cords, and throw pillows so shoppers can mix fabrics and wood stains independently.

Optional hardware & trim packages

Pack optional roof racks, spoilers, or cabinet handles inside one master asset and toggle sub-mesh visibility based on customer selections.

Multi-part mechanical assemblies

Address individual valves, levers, and access panels inside an industrial machine for interactive click inspection and cutaway animations.

Responsibility Tiers

  • 1 Native Aircada layer
  • 1 Aircada Configuration layer
  • 1 Custom Implementation layer
  • 1 External layer

Responsibility Boundaries Across 3D Mesh Splitting

How Aircada's object-set decomposition, Studio sub-mesh wiring, and 3D scene naming divide responsibilities.

Object-set decomposition, viewport splitting & reference-counted GPU cache

  • Native Aircada
  • Single-fetch binary loading with concurrent request deduplication, sub-mesh unit instantiation, interactive Split All/Cut/Duplicate tools, and automatic VRAM disposal when holder counts reach zero.

Sub-mesh material bindings, visibility toggles & transform overrides

  • Aircada Configuration
  • Configured in Aircada Studio by selecting individual extracted sub-meshes in the scene tree and binding material datasets, visibility rules, or animations.

Interactive part-picking menus & zone highlight states

  • Custom Implementation
  • Wiring sub-mesh click triggers to open context-specific UI drawers or highlight the clicked zone on the 3D product.

Logical sub-mesh naming in Blender or CAD

  • External
  • Naming sub-meshes clearly (such as CushionSeat, FrameOak, or Buckle_Left) in your 3D authoring tool so they are easy to identify in Studio.

Frequently Asked Questions

Does splitting a 3D model into 15 sub-meshes cause the browser to download the file 15 times?
No. Through single-payload multi-mesh instancing and Aircada's reference-counted GPU media cache, the browser downloads and parses the source GLB file only once. Concurrent sub-mesh units join the same in-flight load and share parsed geometry buffers across all 15 parts.
Can we hide, recolor, or animate one split part without affecting the rest of the model?
Yes. Because sub-mesh decomposition promotes each extracted part to an independent scene unit while preserving its exact alignment, you can toggle visibility, swap materials, or run keyframe animations on any single sub-mesh. Neighboring parts remain untouched unless you explicitly link them.
What happens in GPU memory when a shopper switches to a different product model?
Aircada releases shared geometry and texture memory automatically once no active scene units reference it. The reference-counted media cache decrements the holder count as each part unloads and disposes the underlying GPU buffers when the final reference drops to zero.
How does mesh splitting interact with the CAD-to-web pipeline?
You can decompose parts automatically during upload or separate additional parts interactively in the viewport afterward. Import settings in the CAD-to-web pipeline preserve your assembly hierarchy on upload, and authors can run Split All, Split Cut, or Duplicate anytime using Studio's Mesh Splitting toolkit.
NEXT STEP

Export configured 3D builds to USDZ, GLB, GLTF, and OBJ

Bake live procedural shaders into portable 2D textures and export customer-configured 3D models for iOS AR Quick Look or CAD workflows.

CONNECTED CAPABILITIES

Related capabilities

Explore the adjacent capabilities and integrations that pair with this workflow.

Engineering CAD to web-ready 3D conversion
CAD TO WEB PIPELINE
CAPABILITY

Engineering CAD to web-ready 3D conversion

Explore
Texture-baked 3D exports to USDZ, GLB, and OBJ
PORTABLE 3D MODEL EXPORT
CAPABILITY

Texture-baked 3D exports to USDZ, GLB, and OBJ

Explore
Interactive 3D part animations and motion playback
KEYFRAME & PROCEDURAL MOTION
CAPABILITY

Interactive 3D part animations and motion playback

Explore
Conditional product rules and part compatibility
CONDITIONAL CONFIGURATION LOGIC
CAPABILITY

Conditional product rules and part compatibility

Explore

Sean Chenoweth
WRITTEN BY
Sean Chenoweth
Co-founder, Aircada
Published Updated