Word and Character Counter

How words are counted

A word here is a run of letters or digits, optionally containing an internal apostrophe or hyphen. So "don't" and "well-known" each count as one word, which matches what most people mean and what most editors report.

Counts vary between tools because the rules vary. Some count hyphenated compounds as two, some count standalone numbers as words and some do not, and some treat an em-dash between words as a separator. Differences of a percent or two between tools are normal and are a definition difference, not an error. If you are held to a specific limit, count with whatever the assessor uses.

Characters, with and without spaces

Both figures are given because different limits mean different things. Social platforms and SMS count every character including spaces. Some academic and publishing limits exclude them. Where a limit matters, check which definition applies before trusting a number.

One subtlety: a character here is a JavaScript string unit, so most emoji and some scripts outside the Basic Multilingual Plane count as two. That is also how several platforms count, including SMS, so the figure is usually the practically relevant one.

Reading time

The default of 238 words per minute comes from a 2019 meta-analysis of reading rate studies, which found that figure for silent reading of English non-fiction by adults. It is a better default than the 200 or 250 that most tools use without citation.

Treat it as an estimate. Actual rate varies enormously with material and reader: dense technical prose is read far more slowly than narrative, and anything with code or equations is slower still.

Sentence and paragraph counts

Sentences are detected by terminal punctuation, which is a heuristic rather than a parse. Abbreviations such as "Dr." and "e.g." inflate the count, and decimal numbers can too. Average sentence length is nonetheless a useful signal: much above 25 words and prose usually reads as heavy.

Paragraphs are counted as blocks separated by a blank line, which is how plain text and Markdown are normally structured.

Word frequency

The frequency list excludes common function words and anything shorter than three letters, so what remains reflects subject matter. It is useful for spotting unintentional repetition in a draft.

Treat it with suspicion as an SEO measure. Keyword density has not been a meaningful ranking factor for many years, and writing to hit a percentage produces worse text. The list is more valuable for the writer's question of whether you have leaned on the same word eleven times.

Questions

Why does this count differ from my word processor?

Word counting rules differ between tools, particularly for hyphenated compounds, standalone numbers and text in headings or footnotes. A difference of a percent or two is normal. Use whichever tool your assessor uses.

Are hyphenated words counted once or twice?

Once. A run of letters containing an internal hyphen or apostrophe, such as well-known or don't, counts as a single word, which matches most editors.

Where does 238 words per minute come from?

A 2019 meta-analysis of reading rate research, which found roughly that figure for adults reading English non-fiction silently. It is an estimate; technical material is read considerably more slowly.

Is my text sent anywhere?

No. Counting happens as you type, entirely in your browser. Nothing is uploaded, which matters for unpublished drafts and confidential documents.

Should I write to a keyword density target?

No. Density has not been a meaningful ranking signal for many years, and writing to a percentage degrades the text. The frequency list is more useful for catching accidental repetition.