GLTF size for AR


we’re trying to implement an AR visualization using this library

We use the GLTF downloaded from ShapeDiver but we are facing a few issues :

  • The GLTF size seems super big as I need to downscale it to 0.1 of it’s size in order to display it properly. I’ve tried to change the AR > Model Size parameter to something in mm instead of m but I have the same result.
  • The model is rendering with a lot of glitch issues but I don’t know if it’s related to the fact that we have to manually downscale it or if it’s something else.

A test link here :
Visit with your phone
Simply click on the bottom right icon to open the AR display.

Does anyone already faced this kind of issues?

Thanks for your help.

Hello @Alan11,

Just out of curiosity, are you aware that we already offer an AR functionality like described here?

Regarding the sizing of the model, as the AR scaling is only applied when the model is converted in our AR process, you would have to apply the scale manually in your case. Here is an example on how to do that. While the transformation in this example is a translation, you can of course change that to a scale.

Let me know if that works for you and if you are facing any further issues!

Cheers, Michael

@MajorMeerkatThe3rd thanks for the details. We were trying to use this lib as it’s supposed to be more advanced in the way they handle all devices / OS tech layers. Also in our case the feature is intended to be part of a “share link” capability that implies many load of a single model and we would like to avoid burning all our credits for a simple preview of the model. That’s why we were trying to use an alternate solution.

We will do a few tests to compare ShapeDiver solution with the other one to first check if we face the same glitch issue in both or only in the other one.


Hello @Alan11,

just out of curiosity, could you expand on the “share link” capability that you were mentioning? Maybe we can help you out with something on that end.

Regarding AR, please let us know about features that you feel are missing or errors that you have with our current workflow. We always look to improve our product.

Cheers, Michael

Our users will be able to share a “view only” version of a project where the model is rendered based on a saved state. The viewers only needs to be able to display, rotate or see in AR the model.
That’s why a simple GLTF is sufficient as there will be no way of modifying the parameters of the model.