Glitch Text Generator

How glitch text is built, and why apps clip or strip it

Glitch text is ordinary letters, each followed by a run of Unicode combining marks. Whether a paste keeps the effect depends on how tall the destination lets a line grow and how many marks it accepts per letter.

The combining mark range behind the effect

Every mark this generator adds comes from one block, Combining Diacritical Marks (U+0300 to U+036F), which holds 112 marks. It sorts them into three pools: 33 that sit above the letter, 40 that sit below it, and 21 overlays, hooks and double-width marks that cross the letter or hang off its edge. The Rising, Falling and Core results each draw from a single pool.

Unicode has more combining ranges (Extended from U+1AB0, Supplement from U+1DC0, marks for symbols from U+20D0), but fewer fonts cover them. Staying inside U+0300 to U+036F gives each mark the best chance of being drawn, not shown as an empty box.

Why tall stacks get clipped

A combining mark has no width of its own. It is drawn on the letter before it, and each further mark on the same side is pushed farther up or down, so the stack grows vertically while the line stays the same length. A chat message may let the stack spill over the lines around it. A name field, button or notification usually has a fixed height and cuts off the rest. Fonts differ too: one stacks marks neatly, another piles them on one spot.

Why marks vanish or use up the character limit

Each mark is a separate code point. At intensity 10 one letter can carry up to 22 marks; at intensity 2, at most 6. A field that counts code points fills up long before the visible letters suggest, and the text is cut mid-stack. Some apps also cap or strip long runs of marks and do not publish the threshold. Unicode's own Stream-Safe Text Format draws its line at 30 marks in a row. Normalization does not delete marks, so when a paste comes back clean, the app filtered it deliberately.

More questions

Is glitch text the same as text corrupted by an encoding error?

No. Garbled text such as é in place of é comes from a file being read with the wrong encoding, and it is an error. Glitch text is valid Unicode: the letters are intact and the marks were added on purpose.

How do I remove the marks and get the plain text back?

Delete every character in the range U+0300 to U+036F. A find-and-replace that supports regular expressions can do it with the pattern \p{M}. Be careful with other languages: the same pattern removes genuine accents and vowel signs.

How many marks can one letter carry?

Unicode sets no hard maximum, so the limit is whatever the app or font tolerates. This generator adds between 2 and 22 marks per letter depending on the intensity.

Can screen readers read glitch text?

Results vary. Some read the base letters and ignore the marks, others announce marks one by one or stumble over the word, and search rarely matches it. Keep names and anything people must find or read aloud in plain letters.