The accepted answer is incorrect. It misses characters like ý and Ý, ć and Ć and many others.

I believe this works, but I haven't fully tested it:

text.match(//[A-ZÀ-Ž]/u);