Amazing work first of all. Now some critiques. I think the flaw here is in the assumption that the source material of a drawing has nearly the expressive power to build a UI that is anything other than a static screen with well understood widgets. And then that impoverished source becoming the source of truth in a workflow that constantly produces new output and obliterates any prior improvements to that output. Finally there’s the question of the quality of the code produced. In my experience automatically generated code needs significant rework to be maintainable. That however seems like an obstacle you are more likely to overcome in the short term as long as it is an area of focus for your product. I think it will have very limited uptake though if the code produced isn’t very good and very close what a skilled UI developer would have produced.