What compression does the tool use?

Geometry: Draco (KHR_draco_mesh_compression) or quantization. Textures: WebP (EXT_texture_webp) by default, or AVIF (EXT_texture_avif). The pipeline is built on gltf-transform, the same library most paid GLB compressors run on top of.

Free GLB compressor · runs in your browser

Shrink 3D models
without uploading them.

Q: Is this GLB compressor really free?

Yes. There is no sign-up, no trial, and no paid tier. The compressor runs entirely in your browser, so there is no server cost for us to pass on to you.

Q: Is my 3D model uploaded anywhere?

No. The file is read, decoded, transformed, and re-encoded entirely inside your browser tab using WebAssembly and the Canvas API. We never see the bytes.

Q: How much smaller does it make my .glb file?

With the default Draco + WebP settings, real-world models typically shrink by 70–98%. A 10 MB GLB will often drop to under 1 MB.

Q: What's the file size limit?

There is no hard limit. Browsers can comfortably handle files up to a few hundred megabytes. If you have a 1 GB GLB, expect to wait and to use lots of RAM.

Q: Which 3D file formats does it support?

Input: .glb, .gltf (embedded), .obj, .stl, .ply, .dae (COLLADA), and .fbx (experimental). Output is always a compressed .glb — the modern web-friendly binary glTF format. OBJ, STL, and PLY have no built-in compression, so converting them to a Draco + WebP GLB is the biggest practical size win you can get without losing geometry.

A free online 3D-model compressor that runs entirely in your browser. Drop a .glb, .gltf, .obj, .stl, .ply, .dae, or .fbx in and get a much smaller .glb back — same Draco geometry + WebP texture pipeline the paid services use. Tested down to 98% smaller on real models. No sign-up, no upload, no limit.

Drop 3D models here, or click to choose

One or many · .glb · .gltf · .obj · .stl · .ply · .dae · .fbx

Geometry Textures Max texture

What is a GLB compressor?

A GLB compressor shrinks the file size of a 3D model in .glb (binary glTF) format — usually by compressing the mesh geometry with Draco and re-encoding textures as WebP — so it loads faster on the web, in AR, and in games. This one is 100% free, requires no sign-up, and runs entirely in your browser, so your model is never uploaded to a server. It also accepts OBJ, STL, PLY, COLLADA, and FBX files and converts them into a compressed GLB.

How the free GLB compressor works

A free GLB compressor with the same pipeline as paid sites.

Draco geometry

Industry-standard mesh compression from Google. Big win on detailed models — often 80 %+ off the geometry payload alone.

WebP textures

Textures get re-encoded as WebP at a sensible quality level and optionally resized. That's where the headline savings come from on real assets.

Nothing leaves your tab

Your model is decoded, transformed, and re-encoded in WebAssembly inside the page. No upload, no server, no log entry.

FAQ

GLB compressor questions, answered.

Is this GLB compressor really free?

Yes. No sign-up, no trial period, no paid tier. Because compression runs in your browser, we don't pay for compute time — so we don't charge for it either.

Is my 3D model uploaded anywhere?

No. The .glb file is read, decoded, compressed, and re-encoded entirely inside your browser tab using WebAssembly (Draco) and the Canvas API (WebP). The bytes never leave your machine.

How much smaller will my .glb file get?

With the default Draco geometry + WebP textures + 1024 px max settings, real-world models typically shrink by 70 %–98 %. A 10 MB scanned mesh often lands under 1 MB.

What's the file size limit?

No hard cap. Most browsers comfortably handle GLBs up to a few hundred megabytes. If you have a 1 GB GLB, expect the tab to use a lot of RAM and to take a while.

Which 3D file formats does it support?

Input: .glb, .gltf (embedded), .obj, .stl, .ply, .dae (COLLADA), and .fbx (experimental). Output is always a compressed .glb — the modern, web-friendly binary glTF format.

OBJ / STL / PLY don't have any built-in compression of their own, so converting them into a Draco + WebP GLB is the biggest practical size win you can get without throwing away geometry. .gltf files that reference separate .bin and texture files can't be picked up through a single file input — only embedded (single-file) glTFs work.

What compression does it use?

Geometry: Draco (KHR_draco_mesh_compression), or quantization, or none. Textures: WebP (EXT_texture_webp) by default, AVIF (EXT_texture_avif) as an option. Built on gltf-transform, the same library most paid GLB compressors run on top of.

Will the compressed file still look correct?

For typical PBR-textured models, yes. Draco at the default quantization is visually lossless for most use cases, and WebP at quality 0.85 is a strong sweet spot. If you see artifacts on a specific model, choose Quantize or Keep as-is for geometry, or bump max texture size up to 2048 px.

Can I use the compressed .glb in Three.js, Babylon.js, or model-viewer?

Yes. The output is standard glTF 2.0 with the Draco and WebP extensions. Three.js needs DRACOLoader wired into GLTFLoader; Babylon.js and <model-viewer> support both extensions out of the box.

The family