Try going to e.g. https://el.wikipedia.org/wiki/%CE%A7%CF%81%CE%BF%CE%BD%CE%BF%CE%BB%CF%8C%CE%B3%CE%B9%CE%BF_%CF%84%CE%BF%CF%85_%CE%BC%CE%B1%CE%BA%CF%81%CE%B9%C

Try going to e.g. https://el.wikipedia.org/wiki/%CE%A7%CF%81%CE%BF%CE%BD%CE%BF%CE%BB%CF%8C%CE%B3%CE%B9%CE%BF_%CF%84%CE%BF%CF%85_%CE%BC%CE%B1%CE%BA%CF%81%CE%B9%CE%BD%CE%BF%CF%8D_%CE%BC%CE%AD%CE%BB%CE%BB%CE%BF%CE%BD%CF%84%CE%BF%CF%82 https://el.wikipedia.org/wiki/%CE%A7%CF%81%CE%BF%CE%BD%CE%BF... When visiting the page (at least on Firefox), the location bar displays the intended characters, but when copying it you get the percent-encoded form. If the reason for percent-encoding was just to make scams more obvious, shouldn't the form shown in the browser interface also be URL-encoded? By the way, it seems like that is a much stronger argument when applied to domian names than when applied to the part of the URL after the slash.