What These Words Mean

A character that is really there in your text but draws nothing you can see. It still counts towards a length limit, it still gets copied, and it still travels with the text when you paste it somewhere else.

Sometimes called an invisible character or a non-printing character.

The character occupies no width at all, so it cannot hold a line open or make a name look blank. It is useful for splitting a word without a visible gap, and poor for anything that needs to look empty.

Programmers call these zero-width characters.

The character is about as wide as pressing the space bar, but it is not filed as blank space. That combination is what makes a name look empty while still being accepted by a box that refuses an empty answer.

Every character has one permanent number, written like U+3164. Two characters can look identical and have different numbers, which is exactly the problem these tools exist to show you.

The number is called a codepoint.

Almost every website removes blank space from the start and the end of what you typed before it saves it. It is the commonest reason a blank name comes back empty.

Runs of spaces become a single space. Usual in names, search boxes, and anything that ends up in a web address.

Some websites run a clean-up when they save what you typed, swapping unusual characters for more common ones. Your character may not disappear so much as quietly become a different one, which is worse in a way, because it looks like it worked.

Programmers call this normalisation, and the strict version is NFKC.

Characters that tell software how to lay text out rather than saying anything themselves. A well-built username box strips this whole family out, which is why some invisible characters survive there and others do not.

Their technical name is format characters.

Invisible marks that tell software which way a line should run. They matter when Arabic or Hebrew is mixed with Latin letters or numbers, and removing them can scramble the order the text reads in. Plain Arabic or Hebrew on its own contains none of them.

An invisible marker sitting after a character to choose which version of it is drawn. It is what decides whether something appears as a plain symbol or as a colour emoji, so removing them can change how a message looks.

An invisible suggestion of where a word could be broken if it reaches the end of a line. It shows as a hyphen only when it is used, and it is a common piece of debris in text copied out of a PDF.

An invisible copy of the ordinary alphabet. A whole sentence can be written in them and sit inside normal-looking text, which is how instructions get hidden in a document for an AI to read. Finding these is the reason the detector exists.

Accents piled on top of one letter until the text spills over the lines around it. Removing long runs of them is safe for most writing, but Arabic and Hebrew genuinely use three and four marks on a letter, so that clean-up will damage them.

The effect is often called Zalgo text.

A character from another alphabet that is drawn almost identically to a Latin one, such as a Cyrillic letter standing in for an English a. Not invisible, but the same kind of problem: text that is not what it appears to be.