Type a name with an accent, a Vietnamese vowel, a Cyrillic letter or an Arabic word into any fancy text generator and you get a strange result: some characters transform and the rest come through untouched. That is not a bug in the tool. It is a limit that every generator of this kind shares, and it is worth understanding before you style a name that is not written in plain English letters.
Last updated: 30 August 2026 · SoShareIT Fancy Text editorial team
What actually happens to a non-English letter
The styled alphabets in Unicode cover twenty-six Latin letters, their capitals, and the ten digits. That is the whole set. There is no styled é, no styled ơ, no styled я, no styled ا — not because anyone excluded them, but because these blocks were encoded for mathematical notation, and mathematics uses the plain Latin alphabet.
So a generator meeting a character outside that set has exactly two options: pass it through unchanged, or drop it. Dropping it would destroy the word, so every sensible tool passes it through — which is why you get a line where the ASCII letters are styled and everything else is not.
This is easy to confirm on our own tool. Type a word with an accent into the generator and watch which characters change. The unstyled ones are the ones with no styled counterpart in the standard.
What to expect, by script
| What you type | What happens | Result |
|---|---|---|
| Plain a–z, A–Z, 0–9 | Fully styled | What you expect |
| Accented Latin (é, ñ, ü, å) | Passed through unchanged | Mixed — some letters styled, some not |
| Vietnamese (ơ, ư, ế, ộ) | Passed through unchanged | Mixed, and often heavily so |
| Cyrillic, Greek | Passed through unchanged | Entirely unstyled |
| Arabic, Hebrew | Passed through unchanged | Unstyled, and may also affect layout direction |
| CJK (Chinese, Japanese, Korean) | Passed through unchanged | Entirely unstyled |
Greek has a partial exception worth knowing: the Mathematical Alphanumeric Symbols block does include styled Greek letters, because mathematics uses them. Most generators do not map to them, so in practice Greek usually comes through plain here too.
Accented Latin: the awkward middle case
This is where the result looks worst, and Vietnamese shows it most clearly because so many of its letters carry marks.
A word like Nguyễn has an N, a g, a u, a y and an n that all style perfectly — and an ễ that cannot. The output is five transformed letters and one plain one sitting in the middle, which reads as a mistake even though every character is correct.
There is a technique that half-works and is worth knowing about, along with its cost. You can decompose an accented letter into its base letter plus a combining mark, style the base, and let the mark reattach. Some tools do this.
The cost is real: the combining mark is often positioned for the plain letter, not the styled one, so it can sit noticeably off-centre or overlap. And you have turned one character into two, which doubles its share of a character limit and makes it more likely to be stripped by a field that removes combining marks. It is a trade, not a fix, and for a name people will read we generally would not take it.
What to do instead
- Style the decoration, not the letters. Put symbols and frames around your name and leave the name itself in its own script. This works in any language, keeps the name readable, and keeps it searchable.
- Use a style that is not an alphabet swap. Zalgo and combining marks attach to whatever letter is there, so they work on any script — including Vietnamese and Cyrillic.
- Accept the mixed look deliberately. Sometimes it reads fine, particularly when only one or two characters are unstyled. Look at it before deciding it is wrong.
- Do not strip the accents to make styling work. Removing marks from a name changes the name. It is not a styling decision.
That last one matters more than the rest. A name is not a decoration, and a tool that quietly rewrote it to look neater would be doing something we would not want done to ours. Our editorial policy is why we pass characters through byte for byte rather than normalising them into something more convenient.
FAQ
Why did only some letters change?
The ones that changed have styled counterparts in Unicode; the others do not. Only basic Latin letters and digits were ever encoded in styled form.
Can you add support for my language?
Not by ourselves — the characters would have to exist in Unicode first, and the styled blocks were encoded for mathematics rather than for writing. We map to what the standard actually contains.
Is there any style that works on Vietnamese or Cyrillic?
Yes — anything based on combining marks rather than letter substitution. Those attach to whatever letter is there, so they work on any script.
Why does my Arabic text jump around when I add symbols?
Arabic is written right to left, and adding neutral characters can change how the line is laid out. The text is stored correctly; only the display order changed.
Should I remove accents so the styling works?
We would not. It changes the spelling of the name to gain an appearance, and anyone searching for the real spelling will no longer find it.