Thanks for elaborating on this.
That CMS part is very interesting. How performant is this? We usually have to handle datafiles containing hundreds of invoice xml datasets. Would fetching comparatively small data snippets from a CMS that runs on the same machine (a server with specs so good that it is idle most of the time) or on a machine in the same network be fast enough?
I am doing something similar by calling a node.js server on the same machine providing Javascript functions the old Javascript implementation in Designer cannot handle. But that is only a single data value so far. This works flawlessly. I wonder what would happen with more complex API calls on that scale described above.