A character count measures text in individual characters: letters, digits, punctuation and, usually, spaces. A word count measures it in words. Use characters when the space is fixed, as in a form field, a post, a text message or a title, and words when the limit is about how much you say, as in an essay, an article or a speech. One cannot be converted reliably into the other, because the length of a word varies with the language, the subject and the writer.

What each one measures

A character count is fine-grained and mechanical. Once you have settled whether spaces and line breaks are included, two tools should give the same number for plain text. It tells you how much room the text takes up.

A word count is coarser and closer to how people think about length. It tracks how much is being said and how long it takes to read or speak. It also depends on a definition of "word", which is less obvious than it sounds.

When to use which

Situation Better measure Why
A text box in an online form or application Characters The field holds a fixed amount of text.
Social posts, bios and ad headlines Characters, by the platform's own rules The limit is enforced by software, character by character.
Text messages Neither exactly SMS is measured in encoded units; use the SMS Character Counter.
Essays, articles and assignments Words The limit is about scope and reading effort, not space.
Speeches, scripts and reading time Words Time estimates are worked out from words per minute.
Translation and editing quotes Whichever the contract names A per-character price also raises the question of whether spaces count.
Text in Chinese, Japanese or Thai Characters Spaces do not mark where words begin and end.

Why the two do not convert

Here are two sentences. Each has exactly 12 words.

We sat by the old oak and ate ripe plums at noon.
Environmental regulators consistently underestimate infrastructure maintenance requirements throughout populated metropolitan transportation corridors.
Words Characters with spaces Characters per word
First sentence 12 49 about 4.1
Second sentence 12 151 about 12.6

The same word count, and three times the characters. A box that holds the second sentence would hold three sentences like the first.

You will see rules of thumb that turn words into characters by multiplying by a fixed number. Any such number is an average taken over one particular body of text, and it moves when the text changes:

  • Language. "Speed limit" is two words and 11 characters in English. The German equivalent, Geschwindigkeitsbegrenzung, is one word and 26 characters.
  • Subject and register. Specialist vocabulary runs long and everyday vocabulary runs short, as the two sentences above show.
  • Spaces. A ratio worked out with spaces included is not the same as one worked out without them.
  • Names, numbers and links. A single web address can be longer than a whole sentence and still count as only a few words.

If you need an estimate, measure your own writing instead of borrowing a ratio. Paste a few paragraphs of the same kind of text into the Character Counter, divide Characters by Words, and use the result for that kind of text only. It is still an estimate. When the real limit is in characters, count characters.

One caution about the tool's own figures: "Average word length" in the Advanced statistics panel counts only the characters inside words. It leaves out spaces and punctuation, so it comes out lower than characters divided by words.

Word counts are not exact either

Ask two tools how many words are in a text and the answers can differ by a few, because "word" has edge cases. Unicode's text segmentation standard calls the interpretation of hyphens "challenging" and the use of the apostrophe "ambiguous". Its default rules keep the hyphen out of words, so it treats "out-of-the-box" as four words, and it leaves implementations free to build compounds on top of that.

This is how the counter on this site handles the usual cases, with the other reasonable answer beside it:

Text Words here Another counter might say
well-known 1 2, if hyphens separate words
state-of-the-art 1 4
don't 1 1 in English; an apostrophe can join two words in French, as in l'objectif
3.14 1 1
1,000 1 1
10 000 2 1, read as a single number
and/or 2 1, if only spaces separate words
one—two (em dash, no spaces) 2 1, if only spaces separate words
🙂 0 1, if anything between two spaces is a word
你好世界 2 1 if only spaces separate words, or 4 if each character counts

The rules behind the middle column: contractions are one word; numbers with a decimal point or a thousands separator are one word; emoji and punctuation are not words; words joined by a single hyphen with no spaces count as one; en dashes and em dashes separate words.

Languages written without spaces

Chinese, Japanese and Thai do not put spaces between words, so a counter cannot find words by looking for gaps. The Unicode standard says that reliable detection of word boundaries in such languages "requires the use of dictionary lookup or other mechanisms". This site uses the browser's built-in segmenter (Intl.Segmenter), which includes dictionaries for that purpose. The result depends on the dictionary your browser ships, so treat a word count in these languages as an estimate and prefer the character count.

Choosing the number to work to

  1. Use the unit the requirement names. Do not check a character limit by converting it to words, or the reverse.
  2. For a word limit you are close to, find out which tool the person checking will use. Hyphenated compounds and numbers can move a count by a few words.
  3. For a character limit, settle the two questions that change the number: whether spaces count, and how line breaks are counted.
  4. Watch both while you edit. The counter shows characters and words side by side, which is useful when a brief gives a word limit and the submission box has a character limit. For cutting text down to either, see writing to a strict character limit.