Handling umlauts in document name, references etc...

I’ve been getting an error when I compile to LaTeX: “Missing \endscname inserted…”, which I’ve traced to a \newlabel command in the .aux file generated via Scrivener’s MMD->LaTeX compile settings, as set up in Ioa’s sample project (now heavily modified by me) in this topic: https://forum.literatureandlatte.com/t/best-approach-for-creating-mixed-layout-document/42112/1

The line in the .aux file that it points to is this one:

\newlabel{l\IeC {"u}tton}{{\M@TitleReference {1}{L\IeC {"u}tton }}{29}{Lütton}{section*.44}{}}

As experienced LaTeX users might be able to see, It’s having trouble interpreting the umlaut in “Lütton”, instead substituting ‘l\IeC {"u}tton’ in the lower-case label, and later ‘L\IeC {"u}tton’ in the capitalized version.

Is there something I can do in Scrivener to make it not throw this kind of error when I use extended characters? I’ve tried adding \usepackage[utf8]{inputenc} to the LaTeX header (though maybe I didn’t put it in the right spot among all the other things in the header?). I think that fixed other errors (can’t recall at this point, I’ve gone through lots of iterations over several days), but I wonder if it’s causing this particular error.

Could it be handled by a transformation, where it simple replaces ‘{l\IeC {"u}tton}’ with ‘{lutton}’ (and similarly with the capitalized version)? Is there a more LaTeXy way of fixing this that would be work for all extended characters?

using XeLaTeX to typeset might solve this quickly. It has UTF-8 support built in, so - – — uü etc works from the .tex document.

Also usually, though not always a LaTeX file can be read/typeset by XeLaTeX any changes.