Symbols to copy and paste
245 results
Arrows
Currency
Maths
Punctuation
Greek
Stars and shapes
Legal and office
Fractions
Box drawing
A static table · nothing is uploaded
242 characters in nine groups: arrows, currency, maths, punctuation, Greek, stars and shapes, legal and office marks, fractions and box drawing. Three of them — π, ° and № — are listed in two groups each, because that is where people look for them. Tap one to copy it. The search box matches the name as well as the character, so typing "euro" finds €, and typing € finds it too.
How to copy a symbol
These are ordinary Unicode characters rather than images, so they behave like letters. They inherit the surrounding font and colour, a search box can find them, and a screen reader announces them by name. An em dash copied from here is a real em dash, and it survives being emailed, stored in a database and read back out.
Several pairs in this table look identical and are not the same character, which matters the moment anything compares two strings. The micro sign µ (U+00B5) sits in the maths group and the Greek letter mu μ (U+03BC) sits in the Greek group; they are different codepoints, so a search for one will not find the other, and uppercasing µ turns it into the Greek capital Μ. The same trap holds for the increment sign ∆ (U+2206) against the Greek capital delta Δ (U+0394), and for the multiplication sign × (U+00D7) against a lowercase letter x. When two strings that look identical refuse to match, this is usually the reason.
Unicode normalisation is the other thing that quietly rewrites these. A compatibility pass, NFKC or NFKD, is meant to fold look-alikes together, and it does so aggressively: ½ becomes the three characters 1, fraction slash, 2; ² becomes a plain 2; № becomes No; and µ becomes the Greek μ. Plenty of systems run that pass without telling you, including some search indexes, username checks and filename handlers. A canonical pass, NFC or NFD, leaves all four alone. If a symbol has to survive a round trip exactly as typed, check which form the system on the far end stores.
Font coverage is the practical limit, not the standard. Every character here exists in Unicode; whether it appears depends on the font in use carrying a glyph, and a font that does not shows a replacement box. Arrows, currency, maths and Greek are close to universal. Box drawing, the rarer fractions and the office marks are patchier, and a symbol that renders correctly in this browser can still box out in a PDF, in a slide deck or on someone else's phone. A fallback font stack fixes it; swapping in a different character does not.
Two absences are deliberate and one is not. The fractions here are the ones Unicode gives a single character to, so ½, ¼ and ⅐ are one codepoint each and there is no character for seven sixteenths; those are built from the digits and the fraction slash. Superscripts stop at ³ because ¹ ² ³ are Latin-1 leftovers and ⁴ onward live in a different block. And there is no true minus sign U+2212 in this table, only the maths operators around it, so a minus in running text is still the hyphen on your keyboard, which is narrower than the plus sign it is meant to line up with.
One case catches people out every time: ½ and ¾ look like numbers and are not. A spreadsheet reads them as text, Number("½") is NaN, and any arithmetic on them fails without an error. Type 0.5 where a value is meant and keep the character for prose.
What people use it for
- Getting a real em dash instead of a hyphen doing the job of one
- Typing a currency sign that is not on the keyboard
- Putting a degree, per-mille or estimated sign into a report
- Adding an arrow to a slide without drawing one
- Finding the × that is a multiplication sign and not the letter x
- Marking feet and inches with primes rather than apostrophes
- Drawing a table or a tree in a plain-text file with box characters
Questions
No, they are real Unicode characters. They take the font and colour of the text around them, and a search box can find them.
242, across nine groups — 245 entries, since π, ° and № each appear under two headings. It is a working set rather than the whole standard, which runs to more than 150,000 characters.
The font in use has no glyph for it. The character itself is intact: add a fallback font rather than choosing a different symbol.
Arrows, currency, maths and Greek. Box drawing, the rarer fractions and the office marks are the ones most likely to box out in a PDF or on a phone.
No. It is a character, not a number, and a spreadsheet reads it as text. Number("½") comes back as NaN. Use 0.5 for arithmetic.
Unicode only assigns single characters to the common fractions. Anything else is written as digits around a fraction slash, U+2044.
¹ ² ³ are inherited from Latin-1 and sit in the same block as the fractions. Superscript four onward lives elsewhere in the standard.
The first is the micro sign, U+00B5, kept for units like µm. The second is the Greek letter mu, U+03BC. They render the same and compare as different.
Almost always a look-alike pair: micro sign against Greek mu, increment against Greek capital delta, or the multiplication sign against the letter x.
A compatibility pass, NFKC or NFKD, rewrites several of these: ½ becomes 1⁄2, ² becomes 2, № becomes No, µ becomes μ. A canonical pass, NFC or NFD, leaves them alone.
No. U+2212 is not in this table, so a minus in prose is still the keyboard hyphen. It is narrower than the plus sign, which shows in aligned columns.
′ and ″ mark feet and inches, and arcminutes and arcseconds. They are not apostrophes and not quotation marks, though most keyboards give you the apostrophe.
Unicode calls it the estimated symbol. It sits next to the net quantity on pre-packed goods sold in Europe and means the printed figure is an average across the batch, not a minimum guaranteed in every pack.
Usually, on any modern filesystem. Windows still refuses the ASCII characters \ / : * ? " < > |, none of which are in this table.
Yes, after percent-encoding, which the browser does when you paste. One symbol costs several characters: € becomes %E2%82%AC.
Yes, by their Unicode name, which is helpful for a currency sign and tedious for a row of box-drawing characters. Keep decorative runs out of anything meant to be read aloud.
Yes, in any UTF-8 column. The one old trap is MySQL's three-byte utf8, which cannot store four-byte characters; nothing in this table needs four bytes, but emoji do.
Each platform has its own default fonts, and the same arrow is drawn with different weights and proportions in each. The character is the same.
Yes. Keep them out of cells that hold values, since a symbol anywhere in a cell turns the whole cell into text.
« » are the quotation marks used in French, Russian and several other languages. English uses “ ” and ‘ ’, which are also here.
Yes, U+203D, a single character combining a question mark and an exclamation mark. Font support is uneven, so try it in the destination before committing to it.
Tables, trees and frames in fixed-width text: terminal output, code comments, README files. It needs a monospaced font to line up.
No. It is a static table and copying happens in your browser.