Bold Text Generator: How Unicode Fonts Work (2026)

What 'bold text' actually is when you copy it

Open the bold text generator above, type the word 'hello', and copy the output. Paste it into a plain text editor that shows code points. What you pasted is not the ASCII string 'hello' with a bold style applied. It is five distinct characters from a Unicode block called Mathematical Alphanumeric Symbols, each one a separate code point that the Unicode Consortium assigned its own name and meaning.

The character that looks like a bold lowercase h is named MATHEMATICAL BOLD SMALL H, code point U+1D421. To a font renderer that is a completely separate character that happens to have a shape resembling a heavier weight of the Latin alphabet. To a regex looking for the letter h, it does not match.

This is why bold Unicode survives platforms that strip formatting. Twitter does not store font weight on a bio. It stores the raw Unicode string. The string can contain any printable code point, including the Mathematical Bold versions of every letter. The visual effect of bold text appears because the characters themselves are shaped that way, not because the platform applied a style.

The Mathematical Alphanumeric Symbols block (U+1D400-U+1D7FF)

Unicode 3.1 introduced this block in March 2001 to support typeset mathematics. Mathematicians had been writing tensor names in bold italic, set names in double-struck capitals, and number-theoretic functions in fraktur for decades, and the PDF and HTML representations of mathematics papers needed code points to encode those distinctions.

The block runs from U+1D400 to U+1D7FF and contains 996 characters across thirteen styles. The seven this tool exposes render well in standard system fonts. The other six are tuned for specific math notation and ship with limited font support outside scientific publishing.

Within the block, each style follows the same internal layout. Capital A through Z occupy 26 consecutive code points. Lowercase a through z occupy the next 26. Digits 0 through 9 sit in a separate run starting at U+1D7CE. Subtract the ASCII code of A from the input character, add the result to the style's capital base, and the matching styled character pops out.

The seven styles, with copy-paste samples

Mathematical Bold. The most useful style by a margin. Heavy weight, clean shapes, supported by every modern system font. Type: 𝐓𝐡𝐞 𝐪𝐮𝐢𝐜𝐤 𝐛𝐫𝐨𝐰𝐧 𝐟𝐨𝐱.

Mathematical Italic. Slanted serif characters. Type: 𝑇ℎ𝑒 𝑞𝑢𝑖𝑐𝑘 𝑏𝑟𝑜𝑤𝑛 𝑓𝑜𝑥. Note that the lowercase italic h is U+210E (Planck constant) rather than the reserved slot inside the math block.

Mathematical Bold Italic. Heavy weight plus slant. Type: 𝑻𝒉𝒆 𝒒𝒖𝒊𝒄𝒌 𝒃𝒓𝒐𝒘𝒏 𝒇𝒐𝒙.

Bold Script. Cursive calligraphic letterforms. Type: 𝓣𝓱𝓮 𝓺𝓾𝓲𝓬𝓴 𝓫𝓻𝓸𝔀𝓷 𝓯𝓸𝔁. The favorite of Instagram aesthetic accounts and anyone wanting a handwritten feel without uploading a custom font.

Fraktur. Blackletter forms inherited from German typesetting tradition. Type: 𝔗𝔥𝔢 𝔮𝔲𝔦𝔠𝔨 𝔟𝔯𝔬𝔴𝔫 𝔣𝔬𝔵. Hard to read at small sizes. Popular in metal band names and gothic aesthetics. A few capitals (C, H, I, R, Z) live outside the math block at older Unicode positions, so the converter patches those in from the Letterlike Symbols block.

Monospace. Fixed-width characters that look like a terminal font. Type: 𝚃𝚑𝚎 𝚚𝚞𝚒𝚌𝚔 𝚋𝚛𝚘𝚠𝚗 𝚏𝚘𝚡. Useful when you want code-style emphasis inside a platform that strips real code blocks, like a Twitter thread about programming.

Double-struck. Outlined or 'blackboard bold' letters. Type: 𝕋𝕙𝕖 𝕢𝕦𝕚𝕔𝕜 𝕓𝕣𝕠𝕨𝕟 𝕗𝕠𝕩. Standard math notation for number sets (ℝ for the reals, ℂ for the complex numbers). Like fraktur, several capitals are patched in from the Letterlike Symbols block.

Platforms where it works

Twitter and X. Tweets, bios, display names, replies, and lists all accept the full Mathematical Alphanumeric block. Display names render reliably across iOS, Android, and the web. Twitter has restricted usernames (the @ handle) to ASCII because Unicode lookalikes were being used to impersonate verified accounts.

Instagram. Bios, captions, comments, story text, and DMs all render the full block. The Instagram app uses a consistent font fallback across iOS and Android, so what you preview is what your readers see. Bold and bold script are the two styles that scan cleanest on the small Instagram canvas.

Discord. Every channel message and DM accepts the characters. Discord also supports real markdown formatting (asterisks for bold, underscores for italic), so the Unicode trick is most useful for styles markdown does not offer, like fraktur or double-struck.

WhatsApp. All chat messages, statuses, and group descriptions accept the characters. WhatsApp has its own asterisk-and-underscore markdown for bold and italic in chat. Use Unicode for the styles WhatsApp does not natively support.

LinkedIn. Posts, headlines, About sections, and comments all render the full block. LinkedIn is unusual in not supporting any built-in text formatting, which makes Unicode bold a common workaround for emphasis inside long-form posts.

Platforms where it breaks

Slack often. Slack messages technically accept Unicode bold, but the workspace search index treats the bold version of a word as a different string from the ASCII version. A message containing 'launch' in bold will not surface when a coworker searches for 'launch' in plain text. Slack also renders the characters in the workspace font, which on many corporate themes does not match the surrounding text weight.

Code editors and config files. Programming languages parse identifiers by code point class. A variable named 𝐌𝐲𝐕𝐚𝐫 is a syntax error in most languages. JSON, YAML, and TOML keys with bold Unicode will fail validation on the consuming side. Embedded Unicode bold inside string values is fine, but never inside an identifier.

Forms. Login forms, signup forms, and most input fields validate input against a restricted character class. Email locals have to be ASCII to round-trip through SMTP. Phone fields and credit card fields are ASCII-only by definition. Pasting bold Unicode into a username field on most websites produces a 'username must contain only letters and numbers' error, even though the bold characters are technically letters in the Unicode sense.

Search. Browser find-in-page and most search engines treat the bold version of a word as a completely different string. A page titled with bold Unicode will not appear when readers search for the plain-text version. This makes Unicode bold a bad choice for headings on any page you want indexed.

The accessibility cost

Most screen readers announce each Unicode character by its formal name unless the reader has a custom pronunciation rule for that character. So the word 'hello' typed as 𝐡𝐞𝐥𝐥𝐨 in Mathematical Bold is announced as 'MATHEMATICAL BOLD SMALL H, MATHEMATICAL BOLD SMALL E, MATHEMATICAL BOLD SMALL L, MATHEMATICAL BOLD SMALL L, MATHEMATICAL BOLD SMALL O' by default. That is unusable for any reader who relies on a screen reader for navigation.

VoiceOver on macOS and iOS, NVDA on Windows, TalkBack on Android, and the screen reader built into ChromeOS all behave this way out of the box. A reader using JAWS or NVDA at default settings will struggle to understand a bio written entirely in bold Unicode.

The practical rule: use real bold for content. Markdown asterisks, HTML strong tags, or rich text formatting all produce semantic bold that announces correctly and survives copy-paste into rich-text destinations. Reserve the Unicode trick for visual flourishes in usernames, bios, and one-off captions where the visual is the entire point and a screen reader would not encounter it as a primary reading surface.

When bold Unicode is the wrong tool

Three contexts where the answer is always 'use real formatting, not Unicode':

Article body content. If you are writing a blog post, a documentation page, or an email newsletter, use markdown or HTML bold. Real bold is one tag away, announces correctly to assistive tech, scales with the reader's font preferences, and copy-pastes into other rich-text destinations cleanly. Bold Unicode in body text is a worse experience for sighted readers (font fallback inconsistency) and a much worse experience for blind readers (literal code-point announcement).

Anything that needs to round-trip through a search index. Page titles, blog headings, product names, anything you want a user to find by typing the plain-text word into a search box. Bold Unicode breaks the search and costs you traffic.

Print or PDF documents. Most PDF readers can render the Mathematical Alphanumeric block, but the font fallback often pulls a different typeface than the surrounding text, which produces a visible weight and shape mismatch on the page. Real bold formatted in the source application produces a consistent print result.

The legal and policy issue

Twitter, Instagram, and Facebook have all updated their username and display-name policies in the last few years to restrict Mathematical Alphanumeric characters. The pattern is the same across all three: bios and post content stay open, but the identity field (the @ handle on Twitter, the username on Instagram, the page name on Facebook) is locked to ASCII or a narrow Latin-extended set.

The reason is impersonation. A verified Twitter account at @nike can be impersonated by an unverified account at @𝐧𝐢𝐤𝐞 (Mathematical Bold). The bold version is a different string with the same visual rendering, so casual readers cannot tell the accounts apart at a glance. Platform trust and safety teams have been closing the loophole one Unicode block at a time. Bios, captions, and tweet bodies are fine to use Unicode fonts in. Usernames and handles are the high-risk territory.

Fun history: the Unicode 3.1 release

The Mathematical Alphanumeric Symbols block landed in Unicode 3.1 in March 2001, alongside the first plane-1 expansion of the standard. Until that release, every Unicode character lived in the Basic Multilingual Plane and fit inside two bytes of UTF-16. Bold, italic, script, and fraktur were among the first non-BMP additions, encoded as four-byte surrogate pairs.

The original use case was scientific publishing: tensor notation in physics, set names in number theory, function families in algebra. For the first decade, almost no one outside scientific typesetting used the characters. Around 2010, internet users discovered that the bold and italic ranges rendered in every modern operating system font and started pasting them into Twitter bios and Tumblr posts. Both communities still share the same code points and inherit the same accessibility problem the standard never anticipated.

Use the tool

Paste any text into the Bold Text Generator, pick a style, copy the output. The conversion runs in your browser. Nothing is uploaded. The output is plain Unicode text you can paste anywhere that accepts text input.

For the case-conversion side of typography (uppercase, lowercase, Title Case, camelCase, snake_case), see the Case Converter Complete Guide. For character-by-character limits on Twitter, Instagram, and SMS, see the Character Count Complete Guide.