Interesting, I don’t know if we’ve tried to load models in a web worker before with a null engine. Seems interesting.
In a web worker, there is no way to get the absolute url from a relative url, so I guess we have to update the code to assume the url is already absolute if document doesn’t exist. Will this work for you? Are you hosting your own Draco compression files and can you point to the absolute url?
As a workaround, maybe try defining something like this before calling ImportMeshAsync to trick it into doing nothing instead of throwing an exception: