Hello,
Scratch 3 has turned out to be a problem because MIT has not provided a mechanism for third parties to integrate (as they previously did with scratchx in scratch 2).
The only way I can see to do it is to take the entire open source scratch source tree and deploy it on my own server. This will increase my expenses quite a bit, though, and I feel like it's going to be painful to get it working.
So I am evaluating other alternative systems that are similar like snap and blockly. But it turns out there are a huge number of such alternatives, meaning it's slow going to evaluate them all and figure out which one to commit to. I also keep hoping MIt will open up their system to outside integrations. Alas, they seem interested more in large corporate integrations than the open source community.
I realize chrome will stop supporting scratch 2 by end of 2020. I should have something up and running before then.