A superscript generator turns your normal text into tiny raised characters like ᵗⁱⁿʸ ᵗᵉˣᵗ that you can copy and paste straight into a bio, caption, or username. It doesn't shrink your font or apply real formatting — it swaps each letter for a separate Unicode character that already looks small and lifted off the line. Subscript does the same thing in the other direction, dropping characters below the baseline (ₐ ₑ ₒ). Because the style is baked into the character itself, it travels with the text into places that normally strip formatting.
That's the good news. The honest news is that neither superscript nor subscript is a complete alphabet, so a handful of letters quietly fall back to normal size — and on some devices, the text shows up as boxes. Here's exactly how it works, real examples you can lift, and where it breaks.
What "tiny text" really is (the Unicode mechanic)
When you bold a word in a document, you're applying a style on top of an ordinary letter. Tiny text doesn't work that way. There is no "make it small" instruction attached. Instead, the generator replaces each character with a different character that was designed, from the start, to render small and raised (or small and lowered).
The lowercase 'a' in ᵃ is not a styled 'a' — it's a distinct code point that just happens to look like a miniature, elevated 'a'. Your phone draws it that way because that's what the character is. This is the same reason BoldlyType's bold and italic Unicode styles survive a copy-paste into an Instagram bio: the appearance is part of the character, not a coat of paint on top of it.
Where do these characters come from? Not one tidy place. They're scattered across several Unicode blocks for historical reasons:
- The main Superscripts and Subscripts block (U+2070–U+209F) holds superscript zero, superscript digits 4–9, all ten subscript digits, and a small set of subscript letters.
- Superscript 1, 2, and 3 aren't even in that block. They live in the older Latin-1 Supplement block (U+00B9, U+00B2, U+00B3), encoded decades ago for legacy compatibility — which is why the superscript digit set is split across two blocks.
- Most of the raised lowercase letters (the actual "tiny text" alphabet) come from the Phonetic Extensions block (U+1D00–U+1D7F), where superscript 'a' sits at U+1D43. These were created for the International Phonetic Alphabet, not for decorating captions.
So when someone says "they all come from the superscripts block," that's not true. The set is a patchwork, assembled from math notation, phonetics, and old compatibility characters. That patchwork origin is the root cause of every gap you're about to hit.
Real copy-paste examples
Here are working examples you can copy directly. Each was built by character substitution — paste them anywhere and the raised or lowered look comes along.
Superscript (raised):
- ᵗⁱⁿʸ ᵗᵉˣᵗ ᶠᵒʳ ᵇⁱᵒˢ
- ᵃᵉˢᵗʰᵉᵗⁱᶜ ⁿᵒᵗᵉˢ
- ʷᵉˡᶜᵒᵐᵉ ᵗᵒ ᵐʸ ᵖᵃᵍᵉ
Superscript digits and math (these are rock-solid because the full digit set exists): x² · 10⁶ · 3ʳᵈ · n⁻¹
Subscript (lowered):
- ʜₑₗₗₒ (note: only the letters that exist render small)
- aₙ · xₖ · vₘₐₓ
Subscript numbers and formulas (fully supported): H₂O · CO₂ · x₁ + x₂ = y · a₀, a₁, a₂
Notice something in the subscript word examples: some letters drop down and some don't. That's not random — it's the coverage gap, and it's worth understanding before you rely on this for anything.
Coverage gaps: where it quietly breaks
This is the part most generator pages skip. Being honest about it is the difference between text that looks intentional and text that looks broken.
Superscript is missing one key letter
For the practical raised alphabet, the one real problem lowercase letter is 'q'. A superscript 'q' is effectively missing from the common, well-supported character set — it only exists as an obscure modifier letter added to a supplementary Unicode block in 2021 (Unicode 14.0), which most fonts don't render, so generators treat it as unavailable and substitute a workaround. Almost every other lowercase letter has a clean, widely supported superscript form — including 'i' (U+2071), which is a normal dotted superscript i, not a broken one.
Uppercase is where the bigger gaps are. The superscript capital row is not a full A–Z — capital Q and Y aren't part of the usable set (Y has been provisionally floated for a future Unicode version, but you can't rely on it today). So "make any word superscript" is a promise no honest tool can keep: words with a lowercase 'q', or a leading capital Q or Y, will either approximate that letter or leave it full-size.
Subscript is far sparser
Subscript looks like the symmetrical twin of superscript, but the lowercase letter set is much smaller. The subscript lowercase letters that actually exist are: a, e, h, i, j, k, l, m, n, o, p, r, s, t, u, v, x — seventeen letters. Missing entirely: b, c, d, f, g, q, w, y, z. And there are no uppercase subscript letters at all.
One quirk worth knowing: most of those subscript letters sit together in the Phonetic Extensions and main Subscripts blocks, but subscript 'j' is the odd one out — it lives by itself at U+2C7C in the Latin Extended-C block. It's real and it works; it's just easy to miss, which is probably why it's sometimes wrongly listed as unavailable.
Even with 'j' included, you genuinely cannot spell most words in subscript — nine lowercase letters and the entire uppercase alphabet are absent. (There are provisional proposals to add subscript w, y, and z, but they're not something to count on yet.) What does work flawlessly is subscript numbers: all digits 0–9, plus +, −, =, and parentheses. So subscript is excellent for math and chemistry notation and disappointing for spelling — the opposite of what people expect.
A note to avoid a common mix-up: the raised "tiny" letters here are superscript modifier letters, which are different from small caps (full-height small letters like ᴀʙᴄ). Both were encoded for phonetic notation, and many of them sit in or near the Phonetic Extensions block — but they aren't all in one place. Small cap A is U+1D00 in Phonetic Extensions, while small cap B and R sit in the separate IPA Extensions block (U+0250–U+02AF). If small caps is the even, on-the-line look you actually want, that's a distinct style — not superscript.
Where tiny text breaks on the reader's end
Coverage gaps explain why some letters don't shrink. A second problem explains why some readers see boxes (□) even for letters that do exist: font coverage. Each tiny character has to be present in the font the viewing app is using. When a device or app lacks that specific glyph, it falls back to a missing-character box. The riskiest characters are exactly the ones pulled from supplementary blocks — like the 2021 modifier 'q' workaround — which many older fonts simply don't include. Older phones, some desktop browsers, and certain in-app text fields are the usual culprits. If your tiny text shows boxes for some viewers, that's a font-coverage gap on their end, not an error in the text.
There are two more limits that have nothing to do with looks. Screen readers handle this text poorly — they may spell it out character by character, read it in a strange voice, or skip it entirely — so anyone using assistive tech may miss what you wrote. And in-app search won't match it: a username or keyword written in tiny text won't come up when someone types the normal-letter version. That's why styled Unicode belongs on decoration, never on the words people need to read aloud, search for, or rely on.
How to use it without it backfiring
Tiny text is a great accent and a poor foundation. A few practical rules keep it on the right side of that line:
- Keep the styled part short. A single raised phrase reads as intentional; a whole paragraph reads as broken on the first device that's missing a glyph.
- Never style anything load-bearing. Handles, links, hashtags, and searchable keywords should stay in plain text so people can find and tap them.
- Lean on numbers for subscript. H₂O, CO₂, and x₁ are rock-solid; subscript words are not.
- Preview on a second device before you commit it to a bio — what looks crisp on your phone may be boxes on an older one.
- Always have a plain-text fallback for anything that has to be read, searched, or read aloud.
On BoldlyType, the conversion runs entirely in your browser — you type, each letter is mapped to its tiny Unicode equivalent, and you copy the result. Nothing is uploaded, there's no signup, and you can preview superscript, subscript, and other styles side by side to see which one survives where you're pasting it. Used as decoration, with plain text underneath it, tiny text does exactly what you want. Asked to carry meaning on its own, it quietly lets you down — and now you know precisely why.
