Bold Text Generator: Six Bold Styles, Digits Included Where Unicode Has Them

Each row below is the same input, "Sale ends 9/30", converted by our engine. The difference in the last column is the one that trips people up: a sale date, a score or a year only turns bold in the first two styles.

The starting code point tells you where the alphabet sits in the Mathematical Alphanumeric Symbols block. A is the first letter of each run, and every other letter follows at a fixed offset, so 𝐁 is U+1D401, 𝐂 is U+1D402 and so on.

Choosing between them is mostly about tone. Bold serif (𝐌𝐚𝐤𝐞 𝐭𝐡𝐢𝐬 𝐛𝐨𝐥𝐝) reads like a heading in a printed book and sits comfortably next to most apps' body text. Bold sans (𝗠𝗮𝗸𝗲 𝘁𝗵𝗶𝘀 𝗯𝗼𝗹𝗱) is flatter and more modern, and it tends to look closest to the bold an app would draw on its own, which makes it the least conspicuous choice for a headline line.

The slanted pair, bold italic (𝑴𝒂𝒌𝒆 𝒕𝒉𝒊𝒔 𝒃𝒐𝒍𝒅) and bold italic sans (𝙈𝙖𝙠𝙚 𝙩𝙝𝙞𝙨 𝙗𝙤𝙡𝙙), work for a single word of emphasis inside a sentence. Bold script (𝓜𝓪𝓴𝓮 𝓽𝓱𝓲𝓼 𝓫𝓸𝓵𝓭) and bold fraktur (𝕸𝖆𝖐𝖊 𝖙𝖍𝖎𝖘 𝖇𝖔𝖑𝖉) are decorative, better suited to a name or a short tag than to anything someone has to read quickly.

The grid shows all 62 plain characters next to their bold serif form. There are no dashed cells and no highlighted cells: bold serif is one of the few styles with a complete, unbroken run of capitals, lowercase and digits, and none of its letters had to be borrowed from the older Letterlike Symbols block.

That completeness is why bold serif is the safest choice for mixed text. Punctuation such as the slash in 9/30, commas, hyphens and the at sign has no bold form in Unicode, so those characters stay exactly as you typed them. In practice the plain punctuation is hard to notice next to bold letters.

If you ever need to check a character by hand, the arithmetic is simple. Bold capitals run from U+1D400 (𝐀) to U+1D419 (𝐙), bold lowercase from U+1D41A (𝐚) to U+1D433 (𝐳), and bold digits from U+1D7CE (𝟎) to U+1D7D7 (𝟗). Any character outside those three ranges in a bold serif line is either punctuation or something the generator got wrong.

Bold sans follows the same pattern from U+1D5D4, with its digits at U+1D7EC. Because neither run has a reserved slot, a bold text generator cannot produce the empty-box problem that affects script, fraktur, double-struck and italic. If a bold serif or bold sans letter shows as a box, the cause is on the viewing device, not in the conversion.

The coverage table is our own measurement from 2026-10-27. We read the character map of each of the 365 system font files on macOS 26.5.1 and of the Android system fonts that carry these blocks, and counted how many of each style's characters at least one font can draw.

All four styles came out complete on both systems. That rules out the most common cause of "my bold turned into boxes" on those platforms. We did not measure Windows, so the table says nothing about it, and an app that ships its own fonts can still behave differently from the system.

Bold works best as a signpost: one short line that tells people what a post is about, or a name that should stand out in a list. It works worst where text needs to be found, typed or read aloud.

On sites where the post box has no bold option, Unicode bold is the only way to get bold-looking letters into a post, which is why people ask how to bold a tweet or a Facebook update. The pasted characters behave like any other text in that box: they copy, edit and delete like other text, and they appear on readers' screens in whatever font their device uses for that block.

For names, the rule of thumb is to style the field people read and leave alone the field people type. A display name in bold sans is harmless. A handle or username, if the field accepts styled letters at all, becomes something nobody can type when they want to mention you.

Real bold is formatting: the letter stays U+0042 and the app draws it heavier. Unicode bold is a different character, 𝐁 at U+1D401, whose official name is MATHEMATICAL BOLD CAPITAL B. Unicode added it for math, where a bold letter can mean a vector and a plain one a number. Styling a post was never the purpose.

That difference has two practical effects. Search and matching: every math bold letter has a compatibility mapping back to its plain letter, so software that normalizes text will match 𝐛𝐨𝐥𝐝 with bold, while software that compares characters directly will not. Accessibility: a screen reader may announce the formal name, read the plain letter, or skip the character, depending on the reader.

The W3C's guidance on info and relationships asks that meaning not depend on presentation alone. If the app gives you a bold button or markdown, use that. Save Unicode bold for places where there is genuinely no other way, and never put the only copy of important information in it.

A practical middle ground on platforms with no formatting at all: bold only the first few words of a post, the part that shows in a feed preview, and write the rest plainly. The eye still catches the bold line, the body stays searchable, and anyone using a screen reader only has to get through a short stretch of unusual characters.

In a document editor, use its own bold command. Pasting Unicode bold into a document gives you characters the spellchecker does not recognize and that will not change if you later switch the paragraph's font.

When bold output misbehaves, the cause is almost always one of these. Work down the list.

Most of these are not bugs in the generator. They follow from what Unicode encodes, what the receiving app does with unusual characters, and which fonts the reader has. Knowing which of the three you are dealing with saves a lot of trial and error with other tools that use exactly the same characters.