Hello!
We use gtlf-transform in a pipeline:
- step:
gtlf-transform webp file.glb file.webp - step:
gtlf-transform dedup file.webp file.dedup - step:
gtlf-transform draco file.dedup final.glb
The final model can be used with babylon.js with absolutely no issues and let me tell you the file sizes are so small that many times I’m wondering wtf is happening and where are those tens of megabytes packed 
An example, from 93.6 to 9.3 MB…
