Font Generator

Seven Unicode alphabets from one box, each labelled with what it could not convert and whether a search that normalises text will still match your ordinary name.

Why one keyword covers three unrelated jobs

Everyday speech calls any change in letter shape a font, so three groups arrive at the same phrase wanting incompatible things. One wants a typeface file to install. One has a poster and wants to know what type it was set in. One wants a bio or display name that looks different from everyone else's.

Only the third lives inside the text itself, which is the only kind of job a web page can do without a font foundry or an image-matching service behind it. Saying so at the top costs a little traffic and saves a lot of wasted time.

Seven alphabets, and where each one stops

None of these is a typeface. Each is a run of separate characters Unicode encoded for its own reasons — mostly mathematics and East Asian typesetting — that happen to look like styled Latin letters. That origin decides the gaps, and the gaps are not where you would guess.

• Bold — Mathematical Bold from U+1D400. All fifty-two letters and all ten digits, no holes.
• Italic — Mathematical Italic from U+1D434. One empty slot: italic small h was already encoded as the Planck constant at U+210E, which this tool substitutes. No italic digits exist.
• Script — eleven unassigned slots, patched here from Letterlike Symbols. The cursive text generator takes that apart properly.
• Monospace — Mathematical Monospace from U+1D670. Complete, letters and digits.
• Small capitals — not an alphabet at all, but twenty-five phonetic letters gathered from three separate blocks. Unicode has no small capital X, so x passes through as a plain letter and is counted in that row's unstyled tally.
• Circled — U+24B6, U+24D0 and U+2460 onward, complete for letters and single digits. The filled version and its one large hole are on the bubble text generator.
• Full width — U+FF01 to U+FF5E, the only style here that covers punctuation as well as letters. The space becomes U+3000 so the spacing stays even.

Everything outside unaccented A–Z, a–z and the digits is untouched. Accented letters, Greek, Cyrillic, Arabic, CJK and emoji have no styled counterparts, so they pass through or are dropped depending on the control above, and each row counts how many characters that was.

The fold-back badge is the one worth reading

Most of these characters carry a compatibility decomposition: a note in the Unicode database saying which plain character they are a variant of. Run them through NFKC normalisation and the styling dissolves. Bold A, circled A and full-width A all become an ordinary A.

That matters because platforms normalise before they index. Where NFKC is applied, a search for your plain name matches the styled one, autocomplete finds you, and a word filter reads straight through the decoration. Six of the seven styles here behave that way.

Small capitals are the exception, for a historical reason. Those letters were encoded to write down speech sounds, not to restyle text, so Unicode records no relationship between them and the capitals they resemble, and nothing folds them back. Avoid that row if you want to stay findable — and note that most services treat deliberate filter evasion as its own rule break.

The badge is not a lookup. On every keystroke the tool normalises its own output and compares it to your normalised input, so passthrough and dropped characters are accounted for in the verdict you see.

Frequently asked questions

Will people still find my profile if I style my name?

Only on the folding styles, and only where the platform normalises. Bold, italic, script, monospace, circled and full-width all decompose back to plain letters, so a normalising service still matches an ordinary search. Small capitals do not. Plenty of services normalise nothing at all, in which case no styled name is findable. Keeping a plain spelling somewhere in the profile is the reliable fix.

Can I use this to change the font on my website?

No. Pasting these characters into your HTML replaces real letters with substitutes, which wrecks search indexing, breaks find-on-page and turns your headings into noise for screen readers. Type on a site is set with the CSS font-family property, plus an @font-face rule for a webfont.

Why is there no small capital X?

Because the row is assembled rather than designed. Each letter was encoded so a linguist could write a particular sound, which is why they are scattered across IPA Extensions, Phonetic Extensions and Latin Extended-D. No transcription convention called for a small capital X. Q was in the same position until Unicode 9.0 added one at U+A7AF.

Why does one style show rectangles when the rest look fine?

Font coverage is decided block by block. A device can carry the mathematical alphabets and lack the phonetic ones, or the reverse, and you get a box wherever the needed block is absent. The rows above render in your own fonts, so a row that looks broken here will look broken anywhere else on this device.

Accessibility, and what this page cannot tell you

The accessibility cost is real and no setting removes it. Screen readers treat these characters inconsistently: some spell a styled word out letter by letter, some announce each character by its full Unicode name, some say nothing at all where the word should be. A styled display name is a name part of your audience cannot hear, so leave anything load-bearing in plain letters.

Everything is computed inside this page: your text never leaves the device and the tool keeps working with the network off, while the ad slot above will be a separate third-party resource that loads on its own once advertising is enabled. What the page cannot do is preview someone else's phone, so test a name you care about before committing to it.