First of all, great library!
We currently use it at https://coreviz.io/ to create vision AI workflows (multi-step editing/tagging/object detection on images) and it works perfectly for our use case.
One thing I would've liked to see while working with flowcraft is built-in workflow visualization. I was able to create React visualizations by reverse engineeing the vue examples in this repo so I'd like to contribute that back, writing this issue as a mental note to do that.