std::string and Upp::String are two completely different classes.

So the question is, why do you even use std::string in U++ application (or the other way around, why don't you stick solely to std::string, if you want to keep usage of U++ low and having that part of source usable without U++ Core too).

If your app is strongly U++ dependent, I would go for Upp::String only (except places where you call some external library, which has input as std::string, at that single point I would convert the String to string, but otherwise I would keep internally everything in String).

I often write clean C++ libraries in TheIDE (to be compiled on phone platforms, so I don't use Upp at all), then I use std::string only.

Your mixed way is certainly possible too, but I think it's not giving you any advantage, just confusing.