Decorated Text (Fancy Text)

Features

• ⚠️ The first line states that these are different code points, not fonts, and what that costs (unreadable, unsearchable, tofu boxes)
• All 18 styles at once: bold, italic, bold italic, script, bold script, fraktur, bold fraktur, double-struck, monospace, four sans-serif variants, circled, circled filled, squared, squared filled and fullwidth
• Optional combining styles: strikethrough, underline, slash-through and Zalgo (light and heavy)
• A restore mode that turns decorated text — combining marks included — back into plain text
• ⚠️ Combining marks are chosen deterministically: the same input always gives the same output, never a different look on each copy
• ⚠️ Uppercase-only styles (circled filled, squared) leave lowercase alone rather than silently uppercasing it
• Chinese, punctuation and emoji pass through untouched; only letters and digits are swapped

How to use

1. Type the text; all 18 styles are listed at once
2. Turn on the combining switch for strikethrough, underline and Zalgo
3. To restore text someone sent you, switch the mode to Restore and paste it in

FAQ

Why should these not be used for anything formal?

Because they are not fonts but different code points: `𝐀` and `A` are two distinct characters. Three real consequences: screen readers cannot read them (or announce “mathematical bold capital A”), so people who rely on them get nothing; search fails — no site search or search engine will find the `abc` inside `𝐚𝐛𝐜`; and they may not render — if the font serving the page lacks those code points, the reader sees boxes. Great for a nickname, bad for a headline, where some of your readers get a row of squares.

Why are there no superscripts or subscripts?

Because those two sets are incomplete in Unicode. There are only about fifteen superscript letters (`ᵃᵇᶜᵈᵉᶠᵍʰⁱʲᵏˡᵐⁿᵒᵖʳˢᵗᵘᵛʷˣʸᶻ`) and `q` has no superscript form at all; subscripts are even scarcer. The result would be half the letters converted and half left alone, which reads as a broken tool. Better to omit it and say why than to ship a half mapping.

Can Restore recognise every style?

It recognises all 18 listed here, plus the combining styles (strikethrough, underline, Zalgo and so on). It works by looking decorated characters up in a reverse table and stripping combining marks. ⚠️ One case is deliberately kept: ordinary accented letters — the `é` in `café` is `e` plus combining U+0301, and the strikethrough mark U+0336 lives in the same block. So the rule is whether the combining mark immediately follows an ASCII letter: our decorations do, real accents do not.

Why is the Zalgo output the same every time?

Because combining marks are chosen by rotating by position, with no randomness. Plenty of tools in this category use `Math.random()`, so the same input produces a different result on every run — and copying it twice makes the user think they copied wrong. Deterministic output is easier to check and, crucially, possible to restore.