Special Symbols

About Special Symbols

The tool has two layers. The top is a curated set grouped by purpose — faces, arrows, maths, currency, checks, whitespace and so on. Below that, whole Unicode blocks (Latin Extended, Greek, Box Drawing, Braille, Emoticons and dozens more) are laid out in code point order. Click any character to copy it to the clipboard.

The hard part of picking symbols is rarely finding one. It is picking the wrong one that looks identical. Two characters that render the same but have different code points are different to search, sorting, spreadsheets, spell checkers and validation code.

Look-alikes that cause real problems

Unicode encodes characters by identity and origin, not by shape, so the same glyph can appear several times. These are the pairs most likely to cause trouble:

• Minus signs: - (U+002D HYPHEN-MINUS), − (U+2212 MINUS SIGN), – (U+2013 EN DASH). Paste −5 with a true minus into a spreadsheet and many will treat it as text, not a number.
• Quotes: ' and " (typewriter) versus ‘ ’ “ ” (typographic). Word processors auto-convert, so code or JSON pasted from a document often breaks on curly quotes.
• Ohm and omega: Ω (U+2126 OHM SIGN) turns into Greek Ω (U+03A9) under NFC normalization; Å (U+212B ANGSTROM SIGN) becomes Å (U+00C5). A database or filesystem that normalizes on save will not keep the code point you chose.
• Micro: µ (U+00B5) versus Greek μ (U+03BC).
• Latin a (U+0061) and Cyrillic а (U+0430); Latin o and Greek ο (U+03BF). Mixing them in a domain name produces a lookalike address — a homograph attack. Unicode Technical Standard #39 publishes the confusables data browsers and registries use to detect these.
• Multiplication: × (U+00D7) versus the letter x.

Compatibility characters: ①, ㎏, Ａ, ½, fi

Circled digits, squared unit symbols, full-width Latin letters, vulgar fractions and ligatures like fi entered Unicode for round-trip compatibility with older East Asian and vendor character sets. NFKC normalization folds them back: ① → 1, ㎏ → kg, Ａ → A, ½ → 1⁄2, fi → fi, ™ → TM.

That folding is applied by many search indexes and by identifier rules (usernames, some programming languages), so finance and finance may match in one system and not in another. Fine for display; best avoided in filenames, IDs and data values.

Emoji or text?

Older symbols such as ♥ (U+2665), ☺ (U+263A) and ✔ (U+2714) can be drawn either as monochrome text or as colour emoji. A following VARIATION SELECTOR-16 (U+FE0F) asks for the emoji style; VS-15 (U+FE0E) asks for text style. If the same heart looks red in a chat app and black in a document, the difference is often that invisible selector. The UTF-8 byte viewer will show it as an extra EF B8 8F.

Every space has a job

The whitespace group is not a set of interchangeable blanks. A no-break space (U+00A0) keeps “10 kg” or “Mr. Smith” on one line. Thin space (U+2009) is the usual typographic choice for the digit-group space that the SI system uses instead of commas or points (10 000). Em and en spaces are fixed widths for typesetting. Zero width space (U+200B) marks an optional line-break point inside long URLs; zero width joiner (U+200D) glues emoji sequences together. Using U+200B to fake an empty username is widely blocked, and it breaks copy-paste for whoever receives your text.

Why some cells show a box

An empty box (often called tofu) or a question mark in the block grid usually means your installed fonts lack that glyph, not that the character is missing. The copied character is still valid and will render wherever a suitable font exists. For important documents, remember your reader's device may have the same gap.

Worked examples

How to use it

1. Choose a category or Unicode block, or type a name in the search box.
2. Click a character to copy it.
3. Paste and check it renders as expected in the destination.

Frequently asked questions

Related guides

• Unicode and UTF-8 Explained: From Mojibake to Emoji