Unicode Text Converter with Character Cost Check

What does a Unicode text converter do?

A Unicode text converter replaces each plain letter with a different Unicode character that looks styled. It does not apply a font. The bold, italic, script, Fraktur, double-struck, sans-serif and monospace letters come from Unicode's Mathematical Alphanumeric Symbols block, which was encoded for maths notation. Unicode's own report on it says these characters "are not intended for use in non-technical text" (Unicode Technical Report #25). Circled, squared and fullwidth letters come from other compatibility blocks, small capitals from the phonetic blocks, and strikethrough and underline are combining marks added after each letter.

Because they are characters, not formatting, they paste anywhere plain text goes: bios, captions, display names and video titles. That portability is why people use them, and it is also the source of every cost below.

How many characters does styled Unicode text cost?

Most styled letters cost 2 characters instead of 1 in UTF-16, the unit Outfeed's publisher counts Instagram, LinkedIn and Threads text in, and 4 bytes instead of 1 in UTF-8. The maths letters sit outside Unicode's Basic Multilingual Plane, so each one takes two UTF-16 code units. Take "Launch day is here": plain, it is 18 UTF-16 units; in bold, "𝐋𝐚𝐮𝐧𝐜𝐡 𝐝𝐚𝐲 𝐢𝐬 𝐡𝐞𝐫𝐞" is 33. Against an Instagram caption's 2,200-character limit, that one short line spends 15 extra characters.

How much that matters depends on how each platform counts. X publishes its method, the open-source twitter-text weighting table, and under it almost every styled letter weighs 2. Bluesky counts graphemes, what a reader sees as one character, so styled letters cost nothing extra there. YouTube caps descriptions in UTF-8 bytes. Instagram, LinkedIn and Threads publish no counting method; Outfeed's publisher counts them in UTF-16 code units. One styled letter, in each unit:

Put differently: a caption written entirely in bold fits about 1,100 characters in Instagram's 2,200, an all-bold X post about 140 of its 280, and an all-bold YouTube description about 1,250 letters in its 5,000 bytes. Every field the tool checks:

Where the platform does not publish its counting method, the figure assumes UTF-16, which is how Outfeed's publisher counts; treat it as a margin, not a guarantee. To check a whole draft against every limit, use the Instagram character counter or, for X, the Twitter character counter. The LinkedIn post character limit guide covers every LinkedIn field, and the social media sizes table covers images and video.

Why are some letters missing from a style?

Because Unicode encoded 24 of the styled letters years earlier, in the Letterlike Symbols block, and left gaps in the maths alphabets rather than encode them twice. A converter that just adds an offset lands on a reserved, empty code point: italic "h" has to be ℎ (U+210E, PLANCK CONSTANT). This tool uses the correct character every time and tells you when it did. A side effect: those letters cost 1 UTF-16 unit, not 2.

• Italic (serif): h → ℎ (U+210E)
• Script: B → ℬ (U+212C), E → ℰ (U+2130), F → ℱ (U+2131), H → ℋ (U+210B), I → ℐ (U+2110), L → ℒ (U+2112), M → ℳ (U+2133), R → ℛ (U+211B), e → ℯ (U+212F), g → ℊ (U+210A), o → ℴ (U+2134)
• Fraktur: C → ℭ (U+212D), H → ℌ (U+210C), I → ℑ (U+2111), R → ℜ (U+211C), Z → ℨ (U+2128)
• Double-struck: C → ℂ (U+2102), H → ℍ (U+210D), N → ℕ (U+2115), P → ℙ (U+2119), Q → ℚ (U+211A), R → ℝ (U+211D), Z → ℤ (U+2124)

Some styles have no digits at all (italic (serif), bold italic (serif), script, bold script, fraktur, bold fraktur, sans-serif italic, sans-serif bold italic), and small capitals have no q or x. The converter keeps those characters plain and lists them under the style, so a styled headline never silently mixes in a wrong look-alike.

Can screen readers and search read styled Unicode text?

Often not as words. Texas Tech University's digital accessibility guide calls these characters "mathematical symbols", advises avoiding them, and shows a screen reader announcing each letter by its maths name (Texas Tech University). The disability charity Scope notes that screen readers "may skip them entirely or read something irrelevant" (Scope for Business). Bold "Hi" carries the Unicode names "mathematical bold capital h, mathematical bold small i". The tool shows these names for whatever you type.

Search has the same problem. A bold H is a different code point from a plain H, so a search index that does not fold compatibility characters back to plain letters treats them as different words. No platform publishes whether its in-app search does that, so a styled name or headline may be harder to find. Unicode's maths report notes they "can be confused with styled text" and excludes them from security-sensitive uses such as domain names. Use a style for one word or a short heading, and keep names, handles and full sentences plain.

How do you convert styled text back to plain text?

Paste it into the converter: it turns styled letters back into plain ones before doing anything else, and shows you the plain version. Under the hood this is Unicode compatibility normalisation, called NFKC (Unicode Standard Annex #15). Every maths, circled, squared and fullwidth letter has a compatibility mapping to its plain letter, so "𝐒𝐚𝐥𝐞 𝓉ℴ𝒹𝒶𝓎 ⓞⓝⓛⓨ" comes back as "Sale today only". Small capitals have no such mapping, so the tool maps them back itself, and it strips strikethrough and underline marks.

It applies NFKC only to styled letters. Run on a whole caption, NFKC would also rewrite things you meant to keep, such as the ™ sign and the ellipsis. Squared letters exist only as capitals, so their lower case cannot be recovered.

From a styled line to a published post

A style is a finishing touch; the post still has to be written, sized for each platform and published. That is what Outfeed AI does. It is a chat-first AI social media manager: you describe the post, it writes it in your brand voice, shows a preview per platform, and schedules and publishes to Instagram, TikTok, LinkedIn, X, Facebook, Threads, YouTube, Pinterest and Bluesky from one conversation. Its publisher checks text limits in code before sending, for example: Outfeed's publisher refuses a LinkedIn post over 3,000 characters, counted in UTF-16 code units. Outfeed's publisher counts Bluesky posts in graphemes, the way Bluesky does, and refuses anything over 300.