Small caps, tiny text and wide text look like three settings on the same dial. They are not. Each comes from a different part of Unicode, encoded for a different purpose, and each has a different set of letters missing. Knowing which block you are drawing from tells you in advance which words will come out wrong, which saves you from discovering it after you have published.
The small capital letterforms were not encoded for decoration. They come mostly from Phonetic Extensions, U+1D00 to U+1D7F, a block added so linguists could write pronunciation notation precisely.
That origin decides what is available. The block contains the small capitals that phonetic transcription needed and nothing else, which is why the coverage is uneven. Most letters are there. A few are not, and the tool fills those with the nearest shape every device can draw, so a word can come out with one letter sitting slightly off among the small capitals.
There is no small capital digit set either, so numbers stay plain.
Despite the gaps, this is one of the better-behaved styles on the site. The characters are letterforms rather than symbols, they sit inside the Basic Multilingual Plane so they cost one storage unit each, and screen readers handle them better than the mathematical alphabets. The small caps generator is the page for it.
The gap people hit first is the letter Q. Phonetic notation never needed a small capital Q, so Unicode went without one until 2018, when ꞯ arrived at U+A7AF in version 11.0. That is recent enough that font support is still uneven, and on many devices it would come out as a box.
X is the other absence, and a permanent one: Unicode has no small capital X at all. That costs less than it sounds, because a lowercase x is already close to the shape and height a small capital X would have.
So every generator has to choose, and this one favours a character every device can draw over the technically exact one. Q becomes ǫ, an o with an ogonek that has been in Unicode since 1993 and passes for a q at small-cap size, and X becomes a plain lowercase x. Neither is a true small capital, so a word containing them sits slightly uneven, but it never shows a full-size letter or a box. The catch is that ǫ is a different letter to a search engine or a screen reader, which matters if the text has to be found or read aloud.
If your text contains one of the missing letters, you have three options. Rephrase to avoid it, accept the substitute, or pick a different style. Bold and monospace both have complete alphabets and complete digit sets.
The small raised letters people call tiny text come from the superscript and subscript ranges, mostly Superscripts and Subscripts at U+2070 to U+209F plus a handful of characters encoded elsewhere.
Those were encoded for mathematical and chemical notation: exponents, ion charges, molecular formulas. That is a much narrower requirement than a full alphabet, and the coverage reflects it.
The subscript range in particular is sparse. Several consonants have no subscript form at all, because no formula needed one. The superscript range is better but still incomplete.
So tiny text is the style most likely to come out as a mixture of raised letters and full-size ones. It works well for short runs of the letters that do exist and badly for arbitrary words. Type the exact text you intend to post into the tiny text generator rather than a sample, because the sample will not tell you about the gap in your own word.
Wide text, sometimes called vaporwave or fullwidth, comes from Halfwidth and Fullwidth Forms at U+FF00 to U+FFEF.
That block exists for compatibility with East Asian text systems, where Latin letters had to occupy the same cell width as an ideograph so columns lined up in a terminal or a printed form. The wide A is not a decorated A. It is an A drawn to fill a full character cell.
Coverage here is the best of the three. The block has a complete Latin alphabet in both cases, complete digits, and most ASCII punctuation. A word converted to fullwidth comes out entirely converted, which is why the wide text generator is more predictable than the other two.
The cost is length on screen rather than in characters. Each letter is one character but occupies roughly twice the horizontal space, so a wide phrase runs out of room in a narrow field long before it runs out of characters.
Worth adding because people group it with these: the circled letters come from Enclosed Alphanumerics at U+2460 to U+24FF.
That block was encoded for list markers and numbered references, which is why its digit coverage is unusual. It includes circled numbers well past nine, because numbered lists needed them, and its letter coverage is complete for both cases.
The negative or filled variants are a partial set encoded separately, so a filled circled alphabet is more likely to have gaps than an outlined one.
Circled letters have the same small-size weakness as double-struck: the ring around each letter closes up when the text is scaled down, so a bubble word turns into a row of dots at profile-name size. Use it where the text is rendered large. The bubble text generator is the page.
This is where the three styles separate sharply, and it decides which ones are safe near a character limit.
Small caps, superscripts, fullwidth and circled letters all sit inside the Basic Multilingual Plane, below U+FFFF. Each costs one storage unit, the same as a plain letter. A field that counts storage rather than characters charges you nothing extra.
The mathematical alphabets that produce bold, italic, script and blackletter sit above U+FFFF and are stored as a surrogate pair, two units each. In a field that counts storage, those halve your effective limit.
So if you are writing to a tight cap, the blocks in this article are the cheap ones. A 30-character display name holds 30 small capitals and only 15 mathematical bold letters. The character counter reports both numbers so you can see which constraint you are near.
Small caps reads as typographic rather than decorated. It works where you want a difference in texture without the text looking styled: a subtitle, a section label, a name that should look considered rather than ornamental. It is also the most accessible of the three.
Tiny text works as an accent on a short run, and poorly as a full phrase, both because of the coverage gaps and because raised letters are hard to read in quantity.
Wide text works when the field is wide and the text is short. A three-word tagline in fullwidth reads as deliberate spacing. A sentence in fullwidth wraps badly and looks like a formatting accident.
None of the three works in a username field, which restricts you to ASCII regardless of which block you draw from.
If what you want from wide text is air between the letters rather than the fullwidth look, there is a cheaper way to get it.
Inserting a separator between letters gives you the spread without changing the letters at all. Your text stays plain ASCII, so it remains searchable, typeable and readable by a screen reader, and you only pay one character per gap.
The generators on this site have a spacing control for exactly this, with a plain space, a dot, a bullet and several ornaments. It runs before the style transform, so you can combine spacing with an alphabet if you want both.
The difference matters near a cap. Fullwidth costs one character per letter and doubles the visual width. Spacing costs one character per gap and leaves your letters intact, which is usually the better trade.
Every style in this article has holes, and the only reliable way to find yours is to type your own text rather than looking at a sample.
The sample in the box is chosen to look good. Your name, your brand, your tagline may contain exactly the letter that is missing, and the failure is silent: no error, no warning, just one full-size letter sitting among the small ones.
So the check is trivial and skipping it is the mistake. Paste the exact string you intend to publish, read it letter by letter, and confirm every character converted. If one did not, you now know before anyone else does.
That takes about ten seconds and it is the difference between a style that looks deliberate and one that looks broken.
One exists, ꞯ at U+A7AF, but only since Unicode 11.0 in 2018, and font support is still uneven. The tool uses ǫ instead, an o with an ogonek that draws on nearly every device and passes for a q at small-cap size. There is no small capital X at all; a lowercase x stands in.
No. Each fullwidth letter is one character and one storage unit. It takes roughly twice the horizontal space on screen, which is a layout cost rather than a character cost.
All of them, relative to bold or script. Small caps, tiny text, wide text and circled letters sit below U+FFFF and cost one storage unit each, where the mathematical alphabets cost two.