Character counter

What counts as one character

A character here is a grapheme cluster — the unit a person would point at and call one character. That sounds obvious until you meet the cases where software disagrees with people:

• Accented letters. é can be stored as one code point or as e plus a combining acute accent. Either way it is one character here.
• Emoji. A single smiling face is usually one code point. A family emoji is several people joined by zero-width joiners, and a flag is two regional indicator letters. All of them count as one.
• Skin tone modifiers. A waving hand plus a tone modifier is one character, not two.
• Line breaks. One character, and Windows-style CRLF endings are normalised so they do not count as two.

The rest of the world is not consistent about this. Twitter/X weights most Latin text at one unit per character but counts CJK characters as two. Microsoft Word counts UTF-16 code units, so an emoji outside the basic plane comes out as two. Database columns declared as VARCHAR(255) may mean 255 bytes, which for UTF-8 text is a much shorter string than 255 characters. If a limit really matters, the only safe test is the system enforcing it.

Characters with and without spaces

Both figures are on screen at once because different institutions ask for different ones. Translation and transcription work is frequently priced per 1,000 characters excluding spaces — a convention inherited from European agency practice. Academic abstracts and most online forms count spaces. As a rough guide, English prose runs about 17 to 18 per cent spaces, so a 2,000-character passage is around 1,650 characters without them.

Limits worth knowing by heart

These are the ones that come up most often when people are counting characters to a target:

• SMS — 160 characters per segment in GSM-7, 70 in UCS-2.
• Meta description — about 155 characters before truncation.
• Title tag — about 60 characters.
• X/Twitter post — 280 weighted units on the free tier.
• Instagram caption — 2,200 characters, truncated in feed after about 125.
• Alt text — no hard cap, but screen readers and search engines both favour something under 125.

The character limits reference has around forty more, kept in one table, with a note on which of them are enforced and which are only advice.

Writing to a character limit

Character limits punish different habits than word limits do. Long words cost the same as short ones in a word count and much more in a character count, so the usual moves reverse: prefer use to utilise, help to facilitate, and cut connectives rather than nouns. Punctuation is worth checking too. An em dash with a space on each side costs three characters; a colon costs one, and often does the same job.

One practical trick for a hard cap: write the version you want first and ignore the counter entirely, then cut. Writing under the cap from the first word tends to produce something that reads as if it were written under a cap, which is a different and worse problem than being forty characters over.

Common questions