Skip to content
Burrows Tools

GLB to Meshy-Compatible Converter

Some exporters shrink a GLB with EXT_meshopt_compression, or store its vertices as integers with KHR_mesh_quantization. Both are legitimate glTF extensions, and both are refused by strict importers — Meshy and several other AI 3D pipelines among them — which usually say nothing more useful than that the upload failed. This converter decodes the model in your browser, turns every quantized or integer vertex attribute back into plain 32-bit floats, and writes a new GLB that requires no extensions at all. It then re-reads its own output and reports what came back: size, mesh, vertex and triangle counts, and whether either blocking extension survived. If that check fails it says so rather than handing you a file that will be rejected again. Draco-compressed geometry and KTX2 textures are decoded too. Nothing is uploaded — the file is read, converted and downloaded inside this browser tab.

Drop your .glb here

Or choose it below. The file is read, converted, and downloaded by this tab — it is never uploaded.

How to use the glb to meshy-compatible converter

  1. Drop your .glb onto the panel, or press Choose GLB and pick it.
  2. Read the input summary: extensionsUsed and extensionsRequired tell you whether the file needs converting at all.
  3. Press Convert to plain GLB and wait while the model is decoded, rewritten and exported.
  4. Check the output panel — it re-parses the converted file and confirms both blocking extensions are gone.
  5. The converted file downloads automatically; press Download again if you need another copy.

Frequently asked questions

What exactly does the conversion change?
It decodes any Meshopt- or Draco-compressed geometry, rewrites every integer or normalized vertex attribute — positions, normals, UVs, colours, joint weights and any custom attribute — as 32-bit floats, copies the index buffer into storage of its own, recomputes bounding volumes, and drops the raw extension objects the loader keeps on each node. The result is exported as a new GLB with custom extensions switched off, so nothing that was removed can be written back.
Will the converted file be larger?
Almost certainly, and often by a lot. Quantization and Meshopt exist to make geometry small; storing the same vertices as full floats undoes that on purpose. The point is compatibility, not size — convert as the last step before an upload that is refusing the compressed file, and keep the original for everything else.
Are materials, textures and animations kept?
Materials and textures are re-exported with the model, and textures larger than 2048 pixels are scaled down to that limit because strict importers commonly cap them. Animation clips are not carried over: the export is the scene, not the clips attached to it. If the file had any, the output panel says how many were left behind rather than letting you find out later.
Is my model uploaded anywhere?
No. Decoding, conversion and export all happen in this browser tab, and the download comes from a blob URL created on your own machine. Nothing is sent to a server, and the tool keeps working with your network disconnected once the page has loaded.
What if the verification check fails?
Then the converted file still lists one of the two blocking extensions and the download is deliberately not started, because it would be refused for exactly the same reason the original was. That normally means the model uses a form of the extension this converter did not decode; open it in the GLB viewer to confirm the geometry survived, and re-export from your original tool with compression switched off.

Related tools