Bold and Italic Text for LinkedIn and X Posts

LinkedIn, X, Instagram and most social platforms accept plain text only. There is no markup, no rich-text editor for posts, and no bold button.

The bold and italic text you see in posts is not formatting at all. It is different characters. Unicode contains a block called Mathematical Alphanumeric Symbols (U+1D400–U+1D7FF) holding complete alphabets that happen to be drawn bold, italic, script or monospace:

• Bold: 𝐀 𝐁 𝐂 — U+1D400 onwards
• Italic: 𝐴 𝐵 𝐶 — U+1D434 onwards
• Bold italic: 𝑨 𝑩 𝑪 — U+1D468 onwards
• Script: 𝒜 ℬ 𝒞 — U+1D49C onwards
• Monospace: 𝙰 𝙱 𝙲 — U+1D670 onwards

A "bold" post substitutes each Latin letter for its mathematical counterpart. The platform stores and displays them as ordinary text, because as far as it is concerned they *are* ordinary text.

These characters exist for a specific purpose: mathematical notation, where bold X and italic *X* denote different quantities and the distinction must survive plain text. The Unicode Consortium's own guidance is explicit that they are intended for mathematical use, not for styling prose. Understanding that they were never designed for this explains every drawback in the next section.

This is the part most tutorials omit, and it is the reason to use the technique sparingly.

Screen readers do not read these characters as letters. Because they are mathematical symbols, assistive technology announces them by their Unicode names or skips them. A post beginning 𝐇𝐢𝐫𝐢𝐧𝐠 𝐍𝐨𝐰 may be announced as a sequence of "mathematical bold capital H, mathematical bold small i…" or as silence. Behaviour varies between JAWS, NVDA and VoiceOver and between versions, but no screen reader reads it as the word "Hiring".

For a recruitment post, that means the people who cannot read it include disabled candidates — an outcome directly at odds with the post's purpose.

Search does not match them. Platform search, and general search engines, index 𝐝𝐞𝐬𝐢𝐠𝐧𝐞𝐫 and designer as different strings. Styling your key term removes it from the searches that would have found it.

Copy and paste breaks downstream. Pasted into a spreadsheet, an ATS or a database, these characters survive as themselves — so a candidate's name styled in a profile becomes unsearchable in the recruiter's system.

Fonts may not cover them. Where a font lacks the glyph the reader sees a replacement box. Coverage is good on current desktop and mobile systems and less reliable on older Android devices and in some email clients.

Some platforms de-rank it. Styled text is a well-known engagement tactic, and platforms have historically adjusted for tactics that inflate visibility.

The defensible use is therefore narrow: a few words, never the whole post, never the first line that carries your searchable keyword, and never someone's name.

Most of what people are trying to achieve with fake bold is achieved better by structure.

Line breaks. LinkedIn collapses single newlines in some contexts; a blank line between short paragraphs is reliable and does more for readability than any character substitution. Posts of two to three lines per paragraph consistently outperform dense blocks.

The first 200 characters. LinkedIn truncates posts behind a "see more" link, and the visible portion determines whether anyone expands it. That is where effort belongs — not on styling, but on a first line that earns the click. Keep your searchable keyword in it, in normal characters.

Genuine Unicode punctuation. Bullets (•), arrows (→), em dashes (—) and checkmarks (✓) are ordinary punctuation characters with proper screen-reader support. They structure a post without the accessibility cost, because they are being used for what they are.

Emoji, sparingly. Emoji have accessible names and are announced correctly. One or two as section markers is a reasonable structural device; a line of them is noise.

Real capitals for emphasis, used rarely. A single capitalised word is read correctly by everything.

If you do use mathematical characters, a reasonable rule is: the post must still make complete sense when every styled character is replaced by its plain equivalent — because for a meaningful number of readers, that is effectively what happens.

The transformation is a codepoint offset. For bold uppercase, each letter maps from its ASCII position to U+1D400 plus the same index:

const BOLD_UPPER = 0x1d400; // 𝐀
const bold = (text) =>
  [...text].replace ? text : [...text].map((ch) => {
    const code = ch.codePointAt(0);
    if (code >= 65 && code <= 90)  return String.fromCodePoint(BOLD_UPPER + code - 65);
    if (code >= 97 && code <= 122) return String.fromCodePoint(BOLD_UPPER + 26 + code - 97);
    return ch;
  }).join('');

Two details matter in any real implementation.

These are astral-plane characters. They live above U+FFFF, so each occupies two UTF-16 code units in JavaScript. "𝐀".length is 2, not 1. Iterating a styled string with a classic for loop over .length splits surrogate pairs and produces corruption. Use [...string] or Array.from, both of which iterate by code point.

The blocks have holes. Unicode had already encoded several italic and script letters elsewhere, so those positions in the mathematical blocks are reserved and unassigned. The italic small letter h is not at the expected offset — it is at U+210E, the Planck constant. Script capitals B, E, F, H, I, L, M, R and e, g, o are similarly scattered across the Letterlike Symbols block. A naive offset calculation produces unassigned codepoints, which render as boxes.

Any converter worth using carries an explicit exception table for these. The OpenTools social media post formatter handles the substitution in the browser, which for draft posts — frequently unannounced news, job moves or hiring plans — keeps the text off third-party servers while you are still editing it.