Remove Numbers from Text

In short Removing numbers deletes every digit from text, with options to remove whole numeric words and clean up the double spaces left behind. That second option is what separates a usable result from a mess: deleting the digits from "we grew 40% in 2026" leaves "we grew % in" with stray spacing, while removing the numeric tokens whole leaves "we grew in". Which you want depends on why you are stripping. Anonymising a document usually means removing whole numbers, because a partial figure is still identifying. Preparing text for a language model or a frequency count usually means the same, since a lone percent sign is noise. Two things to check afterwards. Numbers embedded inside words — product codes, chemical formulas, names like "3M" — are damaged rather than removed, and the result is often wrong in a way that is hard to spot later. And dates written in words survive completely, so a document stripped of digits can still be full of "the third of March". Removing digits is not anonymising, and treating it as such is a real risk.