Word Counter

The most common source of disagreement between two word counters is not a bug in either one — it is a difference in what counts as a "word" at the edges. Whether a hyphenated compound like "well-known" is one word or two, whether a standalone em dash between clauses breaks a word count, and whether a string of digits like "42" counts the same as a word made of letters are all decisions a counting algorithm has to make, and different tools make them differently. This tool follows the convention Microsoft Word and Google Docs use — contractions and hyphenated compounds count as one word each — specifically so a total produced here matches what an editor, a publisher or a word-count-based assignment limit expects.

A naive word counter built by splitting text on spaces works reasonably well for English and most European languages, because words in those languages are conventionally separated by a space. It fails completely on Chinese, Japanese and Thai, none of which use spaces between words at all — a whole paragraph of Chinese text has no spaces in it whatsoever, and a space-splitting counter reports it as a single one-word block regardless of how many actual words it contains.

This tool instead uses the Unicode text segmentation the browser itself implements, which applies language-aware rules for identifying word boundaries rather than assuming a space marks every one. That is why it reports a sensible count for mixed or non-Latin text where a simpler tool would report something close to nonsense.

The 238 words-per-minute figure this tool uses for silent reading comes from a large 2019 meta-analysis of reading-speed studies, and it is a population average rather than a fact about any specific reader — actual reading speed varies enormously by familiarity with the subject matter, text difficulty, and the individual reader, easily by a factor of two in either direction. The number shown is useful as a planning estimate — roughly how long a piece will take a typical reader — not as a precise prediction for any one person.

Speaking time uses a slower, separate rate because spoken delivery is paced differently from silent reading: a presenter reading a script aloud clearly, with normal pacing and pauses, moves at something closer to 130 words per minute, which is why the two figures shown are not simply the same number scaled.

A list of the words used most often in a piece of writing is a practical editing aid rather than a curiosity: writers tend to reach for the same handful of words unconsciously, and a piece that leans heavily on "very," "just," or a particular adjective reads as repetitive in a way that is hard to notice while writing it but obvious once flagged. Seeing the actual frequency count turns a vague sense that something reads repetitively into a specific, actionable list of words worth varying.