GEOMETRY & FILE SIZE OPTIMIZATION

Fast mobile 3D mesh compression for heavy CAD

Raw engineering exports and 3D scans often weigh tens of megabytes and stall mobile browsers before a shopper ever touches the configurator. Aircada compresses 3D models automatically on upload with fallback safety and includes a live Studio GLB optimizer for polygon reduction, seam welding, and instant before-and-after file size and mobile memory telemetry.

Fast mobile 3D mesh compression for heavy CAD

Heavy CAD tessellation and mobile conversion drop-off

Engineering CAD assemblies and photogrammetry scans are built for manufacturing tolerances, not cellular connections. A single raw CAD export can carry hundreds of thousands of redundant triangles, duplicated vertices, and fragmented sub-parts that stall page loads and push mobile browser tabs past memory limits.

Aircada addresses geometry weight at two stages. First, during 3D model upload, the ingestion pipeline consolidates duplicate materials (within a configurable 0.22 color tolerance), applies Smart Merge by Material automatically on assemblies over 150 parts, and exports Draco-compressed GLB files (Level 10 default) with automatic fallback if a non-manifold mesh resists compression.

Second, inside Aircada Studio's Media Optimization panel, teams can fine-tune any 3D part in the browser using polygon simplification, vertex welding, and Draco Edgebreaker or Sequential encoding. Faster first-render times keep mobile shoppers actively configuring on Shopify and headless product pages instead of bouncing before checkout.

OPTIMIZATION IMPACT

Cut redundant geometry without losing surface curvature

Illustrative comparison of an unoptimized engineering CAD export against a simplified, welded, and Draco-compressed production GLB.

BEFORE

Raw CAD export (dense tessellation, heavy mobile payload)

Unoptimized engineering exports ship dense flat-face tessellation, unwelded seam vertices, and uncompressed binary vertex buffers that slow initial mobile downloads.

Dense unoptimized 3D CAD mesh wireframe showing excessive triangles on curved and flat seat surfaces
  • Dense triangle counts across flat and gently curved surfaces inflate GPU VRAM usage

  • Duplicate vertices and unmerged sub-parts multiply mobile draw calls

  • Uncompressed binary buffers stall product-page loads on cellular connections

  • No visibility into per-asset disk size, vertex count, or estimated GPU memory

AFTER

Simplified, welded, and Draco-encoded GLB

Studio polygon simplification, vertex welding, and Draco Edgebreaker compression shrink wire weight and GPU memory while preserving crisp product silhouettes.

Clean simplified and Draco-compressed 3D seat wireframe with smooth contours and lightweight polygon density
  • Configurable Simplify Ratio (0 to 1) and Error Tolerance (0 to 1) preserve bevels and curved contours

  • Vertex Welding merges duplicate vertices cleanly prior to bitstream encoding

  • Draco Edgebreaker or Sequential encoding compresses positions, normals, and UVs for fast CDN delivery

  • Side-by-side Studio telemetry reports exact Disk Size, GPU VRAM, Draw Calls, Vertices, and Primitives deltas

How Aircada shrinks 3D geometry from upload to Studio

THREE-STAGE GEOMETRY PIPELINE

01 / POLYGON SIMPLIFICATION

Tune simplification ratio and error tolerance live in Studio

Open any GLB part in Studio's Media Optimization panel to enable mesh simplification with a target Simplify Ratio slider (0 to 1, default 0.5) and geometric Error Tolerance slider (0 to 1, default 0.01). Coplanar and low-curvature triangles collapse first while preserving curved silhouettes, machined bevels, and UV layout.

02 / WELDING & DRACO ENCODING

Weld duplicate vertices and encode with Draco Edgebreaker

Enable Vertex Welding alongside Google Draco compression using either Edgebreaker (default for maximum compression) or Sequential encoding. Vertex positions, surface normals, and UV coordinates quantize into a compact bitstream that streams rapidly over mobile networks.

03 / LIVE TELEMETRY & SCENE PROPAGATION

Verify exact byte and VRAM savings and refresh every unit

The Studio optimizer displays an attribute-by-attribute delta table covering File Size, GPU VRAM, Draw Calls, Vertices, and Primitives, plus clickable links to every scene unit using that asset. Saving immediately updates every referencing unit in your workspace while keeping the original file path intact for one-click Revert or Clone.

3D mesh compression & optimization controls
Stage / ControlVerified Options & DefaultsOperational & Mobile Impact
Upload-Time CompressionDraco Level 10 default with automatic uncompressed GLB fallback; Smart Merge by Material auto-engaged above 150 parts; 0.22 RGB material deduplication tolerancePrevents fragile CAD exports from failing ingestion while shrinking complex multi-part assemblies on initial upload.
In-Browser Polygon SimplificationSimplify toggle (default on), Simplify Ratio (0 to 1, step 0.01, default 0.5), Error Tolerance (0 to 1, step 0.01, default 0.01)Collapses redundant coplanar faces first while protecting curved product silhouettes, machined bevels, and UV seams.
Vertex Welding & Draco EncodingVertex Welding toggle (default on); Draco Edgebreaker (default) or Sequential encoding; configurable position, normal, texcoord, and color quantizationEliminates duplicate seam vertices and packs geometry into compact binary streams for rapid cellular downloads.
Live Studio TelemetrySide-by-side Original vs. Optimized deltas for File Size, estimated GPU VRAM, Draw Calls, Vertices, Primitives, Meshes Size, and Textures SizeGives teams objective byte and memory proof before publishing to mobile storefronts instead of guessing.
Non-Destructive PropagationOne-click Revert to preserved original file path, Clone as separate asset, and automatic cache clearing + scene-unit refresh across every referencing buildLets teams optimize shared 3D parts safely without breaking material assignments, hierarchy transforms, or animations.

Production safeguards built into Studio optimization

GEOMETRY SAFEGUARDS

Measured Disk Size and GPU VRAM deltas

Compare Original vs. Optimized File Size, estimated GPU VRAM, Draw Calls, Vertices, and Primitives directly inside Studio's Media Optimization panel before publishing.

Non-destructive Revert, Clone, and Draco fallback

Studio stores the original unoptimized asset path so you can revert or clone anytime, while upload-time Draco export automatically falls back to standard binary GLB if an edge-case mesh fails compression.

Live propagation across every scene usage

Optimizing a shared 3D part clears its live cache and refreshes every referencing unit across your project automatically while preserving part bindings and animation tracks.

Browser WASM optimization with zero desktop detours

Run polygon simplification, vertex welding, and Draco encoding directly inside your Studio browser tab without round-tripping assets back through Blender or external CLI tools.

Responsibility Tiers

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

Responsibility Boundaries for Geometry Optimization

How Aircada's upload compression, Studio WASM optimizer, and source 3D exports divide responsibilities.

In-browser WASM mesh simplifier, vertex welding, Draco encoding & upload fallback

  • Native Aircada
  • Client-side WASM simplification, vertex welding, Draco Edgebreaker/Sequential encoding, live Disk Size/GPU VRAM telemetry, automatic scene-unit refresh, and upload-time Draco Level 10 export with safe fallback.

Simplify ratio, error tolerance, Draco method & upload split mode

  • Aircada Configuration
  • Configured in Studio's Upload Settings (Smart Merge by Material, All Meshes, Single) and Media Optimization panel (Simplify Ratio 0 to 1, Error 0 to 1, Weld, Draco method, Revert/Clone).

Per-part visibility rules & conditional sub-mesh loading

  • Custom Implementation
  • Binding optimized GLB parts to option rules or modular attachment points so buyers only load the assemblies relevant to their configuration.

Exporting clean source CAD assemblies or 3D scans

  • External
  • Supplying source 3D geometry from SolidWorks, Rhino, Blender, or industrial 3D scanners prior to upload and Studio optimization.

Frequently Asked Questions

Will polygon simplification ruin my UV unwrap or baked textures?
Moderate simplification with a tight error tolerance (such as Studio's 0.01 default) collapses redundant coplanar faces first so existing 2D texture maps stay aligned. Because Studio preserves the original file path and shows live visual previews with a one-click Revert button, you can inspect seams immediately and back off the ratio if a tight label zone needs extra triangles.
How does Aircada shrink 3D file sizes without making curved products look boxy?
Polygon simplification reduces the actual number of vertices and triangles in the 3D mesh, which directly lowers GPU rendering cost and device memory usage. Draco compression (Edgebreaker or Sequential) encodes the remaining vertex positions, normals, and UVs into a tightly packed bitstream so the file downloads much faster over mobile networks.
Can I optimize one shared 3D part without breaking scenes that already use it?
Yes. When you optimize a part inside Studio's Media Optimization panel, Aircada tracks every unit using that asset, clears the stale mesh cache, and refreshes all linked units automatically while preserving material assignments and hierarchy transforms. You can also click Clone to create a separate optimized copy without touching the original asset.
What happens if an uploaded CAD model contains tricky geometry that fails automated compression?
Aircada's upload pipeline enables Draco compression (Level 10) by default and automatically retries as a standard binary GLB if non-manifold or edge-case geometry throws a compression error. You can also control mesh splitting and material deduplication inside Aircada Studio's 3D Upload Settings and pair lightweight meshes with GPU texture compression.
NEXT STEP

Pair lightweight meshes with GPU-compressed textures

Once your polygon counts are dialed in, convert heavy PNG and JPG material maps into Basis Universal KTX2 textures to keep mobile GPU memory low.

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
Split one 3D model into customizable parts
MULTI-MESH DECOMPOSITION
CAPABILITY

Split one 3D model into customizable parts

Explore
GPU texture compression for fast mobile 4K finishes
GPU VRAM & TEXTURE PIPELINE
CAPABILITY

GPU texture compression for fast mobile 4K finishes

Explore
Mobile Safari VRAM and 3D memory protection
MOBILE STABILITY & BATTERY LIFECYCLE
CAPABILITY

Mobile Safari VRAM and 3D memory protection

Explore

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