GLB → GLTF: unpack a single file into editable assets.
Please wait
GLB files bundle models, textures, and materials into one package, which prevents editing individual parts. This tool unpacks a GLB into separate GLTF files: texture images, geometry data, and material definitions become standalone assets. You can then open the GLTF in modeling software and replace or adjust each element independently, without being limited by the packed format.
DefaultY-up
Determines the 'up' axis of the exported model. Different 3D software uses different conventions. Selecting the wrong system may cause the model to lie on its side.
Convention used by OpenGL, Three.js, and most Web 3D engines. Choose this for models from web or game engines.
Convention used by Blender (default), 3ds Max, and some CAD software. Try this if the model appears sideways.
Default0°
Range-180° – 180°
Rotate the model around the X axis (degrees). Useful for correcting axis misalignment or print orientation.
Default0°
Range-180° – 180°
Rotate the model around the Y axis (degrees). Useful for correcting axis misalignment or print orientation.
Default0°
Range-180° – 180°
Rotate the model around the Z axis (degrees). Useful for correcting axis misalignment or print orientation.
DefaultRoot node transform
Choose how rotation is written to the exported model. Baking into vertices rewrites all vertex coordinates so every export format (including geometry-only formats such as STL and STEP) applies it consistently; root node transform keeps the original vertex coordinates, which is friendlier to animated or skinned models, but formats that ignore node transforms (such as STEP) won't apply the rotation.
Rotation is written into all vertex coordinates and applies consistently to every export format; geometry data is rewritten.
Rotation is written as a root node transform while vertex coordinates stay unchanged; better for skinned or animated models, but formats that ignore node transforms (such as STEP) won't apply it.
DefaultSeparated
Choose how textures and buffers are organized in the output.
Export .gltf, .bin, and texture files separately. Suitable for editing individual resources.
Embed all resources as Base64 data URIs in a single .gltf file. Simpler distribution but larger size.
Drag in the GLB file and wait for it to unpack.
Pick separated export (ZIP with standalone texture files) or embedded export (single JSON file).
Download the GLTF assets: texture images keep their original format so you can edit and repack them.
Separated export downloads a ZIP archive with the main .gltf file, model data (.bin), and separate texture images, ideal for editing resources individually. Embedded export downloads a single .gltf file with all resources inlined, convenient for single-file distribution but larger and less convenient for editing textures. You can switch modes in the conversion settings.
In separated mode, you get a ZIP containing the .gltf file, .bin model data, and texture images (.png/.jpg) linked by relative paths. In embedded mode, you get a single .gltf file with everything inlined, no extra files needed.
Yes, use the GLTF to GLB tool on this site to repack the multiple files into a single GLB file. If you have an embedded single-file GLTF, unpack it to separated resources first, or upload that file directly for repacking.
Texture images keep their original format, typically PNG or JPG. If the original GLB uses compressed textures like KTX2, they are converted to standard PNG so you can edit them easily.
Yes. Geometry, materials, textures, and animations are all kept; the conversion only changes the packaging from a single binary file to separate resources. The only exception: compressed textures such as KTX2 are converted to standard PNG for broad compatibility.
The .gltf file will still load the geometry, but materials will appear without textures, usually gray or default-shaded. Keep the whole unpacked folder together so the texture references resolve.
Take a 50MB PBR model through the full pipeline—Blender export, vertex compression, KTX2 texture conversion, engine loading—and get it under 5MB. Includes a complete automation script and a series cheat sheet.
How many bytes does vertex data actually take? Quantization, MeshOpt, and Draco - which tool fits which scenario? It's not about 'Draco is best,' but 'the right tool for the right job.'
ETC1S or UASTC? How to use toktx and gltf-transform? How to load KTX2 in Three.js / Babylon.js? This article is full of copy-paste-ready commands and code.
Your GLB file is only 10MB, but on your phone it eats 200MB of VRAM and crashes. Let's take a closer look at what's really going on.
