Interesting idea of having "unscreenshottable" text:
https://unscreenshottable.vercel.app/?text=Nice%20Try
that uses animated static to make any individual still-frame pure noise, while the eye can still pick out the image/text (if it doesn't give your brain an aneurysm).
But you can still capture the text by screenshotting two successive images and then using ImageMagic to highlight just the differences:
$ composite -compose difference input1.png input2.png output.png
allowing you to view the screen-captured text with in output.png