Convert Fancy Text Back to Plain Text

Convert fancy text to normal here: paste bold, cursive, gothic, bubble, small caps, strikethrough or glitch text and the plain letters appear instantly, with no button to press. It is for anyone who needs to search, edit, sort or read aloud text that was styled with Unicode, and it runs entirely in your browser.

How to convert fancy text to normal

1. Copy the styled text from the post, profile, chat or file. Select whole words. A combining mark at the very edge of a selection is easy to miss, and it will be left behind if it isn’t copied.
2. Paste it into the box. The plain version appears immediately, with no button to press.
3. Switch on the upside-down option only for flipped text. If the result reads backwards or shows strange letters like ǝ and ɹ, turn it on.
4. Copy the result and check it. Skim for capitals that were lost, accents that were removed and decorations you may want to delete.

The converter also removes invisible characters. Zero-width spaces, joiners and byte-order marks are sometimes added to text so that it slips past filters or breaks a word apart without showing. They are deleted along with the combining marks, which is why a word that looked identical to another can finally match when you compare them.

Examples: before and after

Each row below is the actual output of the converter for the word Hello in that style. The last column shows what NFKC normalisation alone would give, so you can see where the extra lookup earns its place.

Three things to notice. Small caps, squares and look-alikes come back in capitals, because those alphabets were built from capitals or have no lowercase to return to. Cherokee shows a case where NFKC does something, but the wrong thing: it turns ϴ into the Greek capital Θ and leaves the rest alone. And the upside-down line only works if you flip the switch, because without it the converter doesn’t know the order is reversed.

When you need plain text back

Styled Unicode looks great until you need to do something with it:

• Search for it. A styled name in a bio or a game lobby often can’t be found by typing the normal letters. If you see a username like 𝐉𝐚𝐧𝐞 and want to look the person up, convert it and search for Jane.
• Edit it. Correcting one letter in 𝕮𝖔𝖔𝖑 means regenerating the whole thing, unless you convert it back first.
• Read it aloud. Screen readers announce many styled letters by their Unicode names. Converting a styled post before sharing it makes it accessible.
• Reuse it in spreadsheets and CRMs. Names pasted from social profiles into a spreadsheet will not sort alphabetically, because each styled letter sorts by its Unicode position, far away from A to Z. Duplicate checks also miss them, since 𝐉𝐚𝐧𝐞 and Jane are different strings. Unstyle the column first, then deduplicate.
• Moderate it. Community moderators can turn deliberately obscured messages, such as banned words written in bold or circled letters, back into readable words before checking them against a filter.
• Clean up titles and headings. Search engines may not match styled letters to the plain words people type, so keep page titles, headings and meta descriptions in plain text. We have no confirmed statement from Google either way, which is a reason to be careful, not a reason to experiment on pages you care about.

What NFKC is, in plain words

Unicode contains many characters that exist only because some older system had them, such as the fullwidth Ｈ used in East Asian typesetting, the circled Ⓗ and the mathematical bold 𝐇. Unicode marks these as “compatibility” characters: they are the same letter as a plain H, drawn differently. NFKC is a standard recipe that swaps each compatibility character for its plain counterpart. The “K” stands for compatibility and the “C” for composed, which means accented letters are glued back into single characters afterwards.

It is built into every modern language, so many websites run it quietly in the background, for example when comparing usernames. It only knows about characters Unicode has explicitly linked. That is why bold, italic and circled letters fold back, and small caps do not.

How the conversion works

Fancy text is built in three different ways, and the converter reverses each one:

Upside-down text needs one extra step. The letters are flipped and their order is reversed, so the upside-down switch undoes both.

What can’t be recovered

• Original capitalisation in uppercase-only styles (black squares, letter tiles, black bubbles) and in look-alike alphabets.
• Look-alikes the converter has never seen. Our lookup is built from the styles on this site, so a character another site invented for the letter “e” is left as it is. Cherokee and Canadian syllabics are the usual trouble: several of these letters resemble Latin ones, but nothing in Unicode links them, so only the pairs we generate ourselves are reversed.
• Accents. Removing combining marks also removes the accent from é, ñ or ü, so Café becomes Cafe. Convert only the styled part of a text when accents matter.
• Decorations. Frames like ꧁ ꧂ and symbols like ★ are kept, since they may be intentional. Delete them by hand if you don’t want them.

Doing it in code

If you only need to clean up a few strings in your own project, the browser and Node can do the main part with no library:

function toPlain(text) {
  return text
    .normalize('NFKD')        // split accents into base letter + mark
    .replace(/\p{M}/gu, '')   // drop all combining marks
    .normalize('NFKC');       // fold bold, circled, fullwidth, superscript
}

toPlain('𝓗𝓮𝓵𝓵𝓸');  // "Hello"
toPlain('H̶e̶l̶l̶o̶');  // "Hello"
toPlain('ʜᴇʟʟᴏ');  // "ʜᴇʟʟᴏ" - small caps are untouched

These are the same normalisation steps this tool uses, minus the look-alike table, so it will not catch small caps, black squares, Cherokee, Canadian syllabics or the rest of the look-alike scripts. It also strips accents from é and ñ. For a full reverse you need a mapping like the one this tool builds from its own styles.

To go the other way, from plain to styled, use the cool lettering generator. For background on why styled letters behave like different characters, read how fancy text works.

If you only remember one thing about how to convert fancy text to normal, it is to copy whole words, so no combining mark is left behind at the edge of your selection.