Compress GLB Online
Processing entirely on your device — large files may take a moment.
Most exported GLB files carry dead weight: textures embedded twice, materials nothing references, vertices repeated per-face. The optimizer rewrites the file with glTF-transform — deduplication merges identical accessors and images, pruning removes anything no scene uses, and welding merges vertices that share the same position, which is what UV-aware exporters often skip.
The output stays a standard GLB — same scenes, same animations, same PBR materials — just without the fat. Files often shrink 10-50% before you even touch texture compression; the result drops straight into three.js, Unity glTFast, Blender or a <model-viewer> tag.
How to compress GLB
- 1
Select your files
Drag & drop your GLB files onto the area above, or click to browse. You can add multiple files at once for batch conversion.
- 2
Convert instantly
Set the options above and click the button. Processing runs entirely inside your browser using WebAssembly.
- 3
Download results
Download each converted GLB file individually, or grab everything at once as a ZIP archive.
About the GLB format
GLB is the binary form of glTF 2.0 — the format the Khronos Group designed as “JPEG of 3D”. One self-contained file carries geometry, a full scene hierarchy, PBR materials with textures, skeletal animation and morph targets. It is the native asset format of three.js, Unity, Unreal, Blender and WebXR.
Converting to GLB is usually the upgrade path from OBJ/STL: materials become physically based and embedded, and scenes become portable in a single file. One honest caveat: Draco-compressed GLBs save space but many loaders need explicit decoder setup — plain GLB plays everywhere unconditionally.
+ Strengths
- ✓PBR materials, hierarchy and animation in one file
- ✓The web/AR standard — loads natively in three.js
- ✓Compact binary with texture embedding
! Limitations
- −Mesh-only: no CAD solids or NURBS
- −Older niche tools may not know glTF
Frequently asked questions
Will the model look different after optimizing?
No. Welding only merges vertices at identical positions, and every mesh, animation and material is preserved. Visual changes are limited to removing things that were never rendered anyway.
Why is my GLB still large after this?
Texture data dominates most large GLBs. This tool removes structural waste but does not re-encode textures — for those, export with smaller texture images or a KTX2 pipeline.
Are Draco/Meshopt compressed GLBs supported?
Files already compressed with Draco or Meshopt need their decoder extensions, which this browser tool does not bundle — export uncompressed input for best results.
Is there a file size limit?
Files up to 50MB process comfortably in the browser; the whole optimization runs on your device, so nothing is uploaded.