A style changes characters, not the font setting
When you choose Serif Bold or Circled, the generator substitutes code points from mathematical, enclosed or width-related Unicode blocks. The destination receives those characters, not a font file or a hidden formatting instruction.
This is why the result can travel through a plain text field. It is also why it may be searched, sorted or spoken differently from the original.
Mathematical alphabets have a specific purpose
Mathematical alphanumeric symbols distinguish variables in technical notation. Their use as decorative text is a practical side effect, not a promise that every app treats them like visual formatting.
Several alphabets contain historical gaps whose characters live in the Letterlike Symbols block. Our registry maps those exceptions explicitly and leaves unsupported characters unchanged.
Combining marks attach to what comes before
Underline, overline and glitch effects append marks that a renderer positions around a base character. Font metrics control the final placement, so marks can overlap, drift or stack differently after pasting.
The Glitch removal preview uses the current variation’s exact source: it removes only marks added here and restores the original input byte for byte. Marks already present in the input remain unchanged.
A compatibility note is evidence, not a score
No single percentage can predict every device, font and app version. We instead document the characters a transformation maps, whether it uses combining marks and whether the mapping can be reversed.
Before using decorative text in a profile, test the exact destination. A box, empty space or replacement symbol usually means the available font lacks that glyph.
- Keep names, links and instructions in ordinary text when discovery matters.
- Paste into the target app and then copy it back to check for filtering.
- Retest after a major OS, browser or app update.
Accessibility needs a meaningful fallback
A screen reader may announce “mathematical bold capital A” instead of simply “A”. Character-by-character names can turn a short decorative label into a long, confusing utterance.
Decorative Unicode does not replace semantic HTML such as headings, strong emphasis or labels. Put meaning in normal text and treat the styled version as optional presentation.
How we verify the catalog
Every style is defined in one registry with its mapping, coverage, Unicode blocks, combining behavior, reversibility and test vectors. Automated tests cover ASCII, accents, joined emoji, right-to-left text, CJK, Devanagari, line breaks and empty input.
We review mappings against Unicode character data and record rendering observations separately from guarantees. Compatibility reports are corrected in the registry first so the tool, interface and tests stay aligned.