Models load faster.
Automatically.
Every model imported into Trice is compressed, quantized, and optimized before you even see it. One-click optimization takes it further — up to 95% smaller files with zero visible quality loss.
Three-stage optimization pipeline
From silent import compression to one-click deep optimization — every stage makes your models lighter and faster.
Import-time compression
Runs silently on every model you import. No buttons, no settings, no waiting. Your model is already optimized before you see it in the viewport.
- Meshopt vertex reordering for GPU cache efficiency
- Quantization — reduce precision where it's invisible
- Binary stream encoding (EXT_meshopt_compression)
Client-side optimization
One-click deep optimization running entirely in your browser. Everything processes in the background — the UI never freezes, even on 200 MB files.
- Mesh: weld vertices, quantize, meshopt compress
- Textures: automatic high-resolution resize
- Textures: convert to WebP for smaller file size
- Real-time progress with before/after comparison
Server-side optimization
Cloud pipeline with native C++ processing. Higher quality compression and operations that browsers simply can't perform efficiently.
- Native sharp library — superior texture compression (quality: 80)
- Geometry deduplication across the entire scene
- Full meshopt pipeline with server-grade processing
- Handles massive models that would choke a browser
- Results identical to local — just faster and better
Mesh merging
Objects that share the same material get merged into a single draw call. A scene with 2,000+ objects can collapse to 5 draw calls — same visual result, dramatically faster rendering.
- Groups objects by material automatically
- Merges geometry into combined meshes
- Reduces draw calls by orders of magnitude
- Massive FPS boost on complex scenes
- Especially impactful on mobile devices
See the numbers
Real optimization results from real models. Same visual quality — dramatically smaller files.
Mesh compression
Vertex reordering, quantization, and binary encoding reduce geometry data while keeping every polygon intact.
Texture optimization
Textures resized to 2048px max and converted to WebP. Massive savings with no visible difference on screen.
Full optimization
Mesh + textures combined. The maximum reduction — models that were 150 MB can drop to under 10 MB.
Why optimization matters for 3D on the web
53% of users leave
If a page takes over 3 seconds to load. Unoptimized 3D models are the biggest bottleneck — a single 100 MB GLB can take 30+ seconds on a mobile connection.
of traffic is mobile
These devices have limited bandwidth, memory, and GPU power. Optimization isn't optional — it's the difference between a smooth experience and a crashed tab.
Textures dominate
In most 3D models, textures account for 80% or more of the file size. Resizing oversized textures and converting to modern formats like WebP is the single highest-impact optimization you can make.
Free vs Pro optimization
Both tiers give you industry-grade compression. Pro unlocks server-side processing for maximum quality.
| Feature | Free | Pro |
|---|---|---|
| Import-time meshopt compression | ✓ | ✓ |
| Vertex reordering & quantization | ✓ | ✓ |
| Vertex welding | ✓ | ✓ |
| EXT_meshopt_compression | ✓ | ✓ |
| Texture resizing | ✓ | ✓ |
| WebP texture conversion | ✓ | ✓ |
| Background Web Worker processing | ✓ | ✓ |
| Before/after comparison | ✓ | ✓ |
| Mesh merging (reduce draw calls) | ✓ | ✓ |
| Geometry deduplication | — | ✓ |
| Quality-controlled texture compression | — | ✓ |
| Native C++ processing (sharp) | — | ✓ |
| Server-grade processing power | — | ✓ |
Frequently asked questions
No. Meshopt compression is lossless for geometry — your model looks identical at a fraction of the file size. Texture optimization uses high-quality WebP encoding which is visually indistinguishable from the original in virtually all cases.
Free optimization runs entirely in your browser using WebAssembly. It handles mesh compression and basic texture resizing. Pro adds server-side processing with native C++ libraries — this means higher-quality texture compression with fine-tuned quality controls, plus geometry deduplication that the browser can't efficiently perform.
Results vary by model. Mesh compression typically reduces geometry data by 60-90%. Texture optimization (resizing + WebP conversion) can reduce texture data by 70-95%. Combined, we've seen models go from 150 MB to under 10 MB with no visible quality loss.
All animations are fully preserved. For rigged models, the optimizer automatically skips certain steps to avoid any chance of breaking bone influences. Your animations play back exactly as before — just in a smaller file.
If you notice anything broken or behaving differently after import or optimization, let us know and we'll look into it. You can also use Raw Import in the Studio menu to skip import-time compression entirely.
All of them. GLB, GLTF, FBX, OBJ, STL, DAE, 3DM, PLY — Trice converts everything to GLB automatically, then applies optimization. Every format benefits.
Yes. Meshopt compression improves GPU vertex cache utilization through reordering, which directly speeds up rendering. Smaller textures mean less VRAM usage and faster texture sampling. Mesh merging can dramatically reduce draw calls — fewer draw calls means higher FPS, especially on mobile devices.
Try it with your own model
Drop any 3D file into Trice Studio and watch the optimizer shrink it instantly. Free, no account required for the viewer.