About Counting Words and Characters | Segmentation, Graphemes and Limits

What is a word?

In English a word is roughly what sits between spaces, but Chinese, Japanese and Thai are written without spaces. The Unicode standard (UAX 29) describes word boundaries, and browsers add dictionaries for those languages. Numbers such as 3.5 count as one word and words with an apostrophe stay together.

What is a character?

A computer stores text as code points, but a reader sees graphemes. A flag, a family emoji or an e with a separate accent mark are several code points yet one visible character. This counter counts graphemes, which is also what most people expect.

Reading and speaking speed

A 2019 review of 190 studies by Marc Brysbaert found an average silent reading speed of about 238 words per minute for English non-fiction, and around 260 for fiction. Presentations and audiobooks usually run at 130 to 160 words per minute. Use your own numbers if you know them.

Common limits

An SMS holds 160 characters from the GSM alphabet, or 70 when it contains other characters, and longer messages are split into parts of 153 or 67. X allows 280 weighted characters. Search engines show about 50 to 60 characters of a title and about 155 to 160 of a meta description, measured in pixels, so treat those as guides.

Frequently asked questions

Do spaces count as characters?

Yes, in the usual character count. The count without spaces leaves out spaces, tabs and line breaks, which some publishers and translators use for pricing.

How are words counted in Chinese or Japanese?

With dictionary based segmentation, so a sentence is split into words of one or more characters. Results can vary slightly between browsers. Many publishers count characters instead of words for these languages.

What is keyword density?

The share of all words taken by one word. Ignoring common words such as the, and or of shows the words that carry the topic. Very high density for one keyword reads unnaturally.