How Unicode Fonts Work

1. A character is a number, a font is a drawing

Every character your phone stores is a number from the Unicode standard: A is U+0041, é is U+00E9, 😀 is U+1F600. A font is a set of drawings for those numbers. When you press "bold" in a word processor, the numbers stay the same and the app stores a formatting note on top. Copy that into Instagram and the note is dropped, because Instagram bios hold plain numbers, nothing else.

The trick behind every fancy text generator: Unicode also contains separate numbers for letters that already look bold, italic or cursive. They were added in Unicode 3.1 (2001) for mathematics, where bold 𝐱 and italic 𝑥 mean different things. Paste U+1D400 and the other app draws 𝐀 because that is what that number looks like.

2. Why some alphabets have holes

Before 2001, Unicode already had a handful of styled letters in its Letterlike Symbols block: ℎ (Planck constant), ℂ for complex numbers, ℝ for real numbers, ℋ for Hamiltonians. Unicode never encodes the same character twice, so the new math alphabets left reserved gaps where those letters would go. Generators have to stitch the older ones back in. That’s why ℯ and ℴ in light script can look slightly different in some fonts: they come from a different part of the font file.

3. Combining marks: how glitch and cursed text grow

Some characters draw nothing on their own and attach to the one before. That’s how é can be written as e + U+0301 COMBINING ACUTE ACCENT. Unicode puts no limit on how many you stack. Strikethrough adds one long stroke per letter. Cursed text adds dozens, above, through and below.

Try it on the cursed text generator, which has a step-by-step stacker. Or the gentler glitch text generator.

4. Look-alikes from other scripts

Styles like "Faux Cyrillic" (НЄLLФ) or "Brushstroke" (卄乇ㄥㄥㄖ) pick characters from other writing systems purely for their shape. They look fun, but they’re real letters in other languages, so they read as nonsense to anyone who speaks those languages and to screen readers. Security teams also watch for look-alikes, because the same trick is used to fake web addresses (a "homograph attack").

5. The cost: accessibility and search

Because styled letters are different characters, software that works with meaning (search, screen readers, translation, spam filters) doesn’t see your word. Try the preview below to hear the difference.

A good rule: style a word, not a paragraph. Keep hashtags, keywords and anything important in plain text. Strikethrough and underline are the safest styles because the original letters stay in place.

This page teaches through small hands-on demos because it’s easier to see a code point than to read about one. You can learn this way about far more on ahaboo, a free site of hands-on explainers built on trusted sources.

Further reading

• The Unicode Standard, chapter 22 “Symbols”, section on Mathematical Alphanumeric Symbols (unicode.org).
• Unicode Technical Report #25, “Unicode Support for Mathematics”, which explains why the math alphabets exist and why styled letters shouldn’t replace normal text.
• Unicode Character Encoding Stability Policy, which guarantees characters are never removed.