What is a zero-width space?

A zero-width space is the Unicode character U+200B, a character that renders as nothing and takes no width. Text editors allow line breaks at it; readers cannot see it. It appears in text copied from web pages and in some machine-generated text, and because it is invisible it can carry hidden marks. Cleaning it out is the first layer of shipping text clean.

The family

U+200B is one of a family of invisible characters: the zero-width joiner and non-joiner (U+200D, U+200C), the word joiner (U+2060), the byte-order mark (U+FEFF), variation selectors (U+FE00 to U+FE0F and the supplementary range) and Unicode tag characters. All of them are legitimate in some script or format; none of them belongs in a paragraph of English prose that you wrote.

Why it matters

Invisible characters survive copy and paste, pass through most editors unseen, and can encode information: a sequence of them is a message a human cannot read. Removing them is verifiable: a scan either finds zero or it does not. On every publication since 2026-08-18 our own pipeline logs that count, and it is zero.

How to see them

Most code editors can render invisible characters (Visual Studio Code highlights them by default). A one-line check in any language counts the code points listed above; our engine does that for text, Markdown, DOCX and PDF from the command line.