Invisible text: copy characters that show nothing

What we verified, on 6 October 2026 against Unicode 16.0: none of the invisible characters here counts as white space, so code that trims spaces leaves them in place, and every one except the Braille blank is default-ignorable, which allows software to draw it with no width at all. We haven’t tested each character in every app. Instagram and Facebook don’t document how they treat them, and an app update can change it.

Many emoji are several characters stuck together by invisible ones. 👩‍💻 is 👩, a zero-width joiner (U+200D) and 💻. ❤️‍🔥 is ❤, a variation selector (U+FE0F) that asks for the emoji style, a joiner and 🔥. If a phone doesn’t know the combination, or the joiner gets lost while copying, you see the pieces side by side: 👩💻 instead of a woman at a laptop.

The flags of England, Scotland and Wales use another trick. 🏴󠁧󠁢󠁳󠁣󠁴󠁿 is a black flag followed by invisible tag characters that spell gbsct, then a closing tag. A phone that doesn’t support those flags ignores the tags and shows a plain black flag.

• Zero-width space (U+200B): a place where a line may break, with no visible gap. Handy inside a very long link or word, and a common stowaway in copied text.
• Word joiner (U+2060): the opposite. It stops a line breaking between the characters on either side of it.
• Zero-width non-joiner (U+200C): stops two letters joining up. Persian uses it in everyday spelling, and Indian scripts use it to control how consonants combine.
• Zero-width joiner (U+200D): makes characters join that otherwise wouldn’t, from emoji sequences to particular letter shapes in Indian scripts.
• Soft hyphen (U+00AD): stays hidden unless the word breaks at the end of a line, where it appears as a hyphen.
• Combining grapheme joiner (U+034F): despite the name, it joins nothing on screen. It exists to keep accent marks in a set order and is otherwise ignored.
• Mongolian vowel separator (U+180E): used in Mongolian script. Unicode classed it as a space until 2013, so some older software still treats it as one.

Arabic, Hebrew and Urdu run right to left, while English and numbers run left to right, and a phone works out the order of a mixed line by itself. Punctuation, brackets and emoji belong to neither direction, so they often land on the wrong side: a full stop at the start of a line, or a bracket facing the wrong way.

1. Spot the character that has jumped, usually a full stop, bracket, colon or emoji at the edge of a phrase.
2. Paste a right-to-left mark (U+200F) straight after it if the line is mostly right-to-left, or a left-to-right mark (U+200E) if it is mostly left-to-right.
3. If that doesn’t settle it, wrap the stray phrase in an isolate: U+2068 before it and U+2069 after it.
4. Save, then check the line on another phone, because apps can lay out mixed text slightly differently.

• Codes and passwords: a zero-width space copied along with a discount code or password makes it a different string, so it fails even though it looks right. Retype it instead of pasting.
• Hashtags and mentions: an invisible character inside #summer or @name can cut the tag short, so it points somewhere else or nowhere.
• Search: a name with a hidden character in the middle may not match when someone types it the normal way.
• Filters: slipping zero-width characters into banned words to dodge filters is a known spam trick, and platforms look out for it.
• Overrides: we left out U+202D and U+202E, which force text into one direction whatever the letters are. The right-to-left one is used to disguise file names, such as making a program look like a photo.