Word Counter

Text is split on whitespace, and any resulting token containing at least one letter or number counts as a word. This means hyphenated words such as "well-known" count as one, contractions such as "don’t" count as one, and standalone punctuation such as a dash on its own line is not counted. It matches the convention used by word processors.

Two reasons. This tool counts Unicode code points, so an emoji or a flag character counts as one character where Word may count it as two. It also counts every character you can see, including newlines. For plain prose without special characters the two figures agree.

Accurate for ordinary prose. Sentences are detected by terminal punctuation, with common abbreviations, initials and decimal numbers excluded so that "Dr. Smith paid $1.50." counts as one sentence rather than three. Unusual formatting — bulleted fragments without punctuation, or heavy use of ellipses — can still shift the number by one or two.

Reading time assumes 238 words per minute, the average adult silent reading rate for non-technical prose found in reading research. Technical or unfamiliar material is slower in practice, so treat the figure as a floor rather than a promise. Speaking time uses 140 words per minute, roughly the pace of a well-delivered presentation.

No. Everything is computed in your browser and nothing is transmitted or stored. Closing the tab discards the text, so keep your own copy of anything important.

The character counts are correct for any language. The word count relies on whitespace, so for Chinese, Japanese and Thai — which do not separate words with spaces — the character count is the meaningful figure to use.