The core rule: words are letter-and-digit runs
This tool counts a word as a run of Latin letters and digits (A–Z, a–z, 0–9). 'The cat sat' is three words; extra spaces never create phantom words. Apostrophes inside a word stay attached, so 'don't' counts as one word.
This is not the space rule: punctuation splits words too. 'well-known' counts as two words, not one. Numbers count as words as well — '2026' is a single word.
What is not a word
Accented letters, emoji, and standalone symbols are not counted as words — only plain A–Z and 0–9 sequences are. Chinese, Japanese, and Korean text contributes zero words; it is measured separately by the CJK character count.
Because of these rules, totals can differ slightly from word processors like Microsoft Word on edge cases (hyphenated compounds, apostrophes). For an exact submission, trust this tool's numbers, not a guess from another program.
Characters: with spaces vs without
Character counts come in two standard flavors. 'With spaces' counts every character including the spaces between words — this is what most platform limits (bios, posts, meta descriptions) mean. 'Without spaces' counts only the letters and punctuation.
The gap between the two is roughly 15-20% for typical prose. Know which flavor your limit uses before trimming: cutting 200 characters to meet a with-spaces limit when the requirement was without-spaces wastes real editing effort.
CJK and non-spaced languages
Chinese, Japanese, and Korean do not use spaces between words, so 'word count' works differently: counters typically report character counts for CJK text instead. A 1000-character Chinese passage is the meaningful unit — not a word count derived from spaces that do not exist.
Mixed-language text follows each part's rule: spaced languages count by words, CJK stretches by characters. If your work spans languages, decide which metric governs each part rather than forcing one number on everything.
Why the rules matter
Understanding the rules turns counting from magic into method. When a platform says '300 words' and your counter says 298, you know the two-word gap is an edge-case difference — not an error — and you can adjust with confidence instead of anxiety.
It also settles disputes: two people with different counts for the same text are usually counting different things (words vs characters, with vs without spaces). Name the metric and the disagreement evaporates.
Frequently asked questions
What counts as a word?
A run of characters between spaces. Hyphenated compounds count as one word; numbers and single letters count as words too.
Why do tools sometimes disagree?
Edge cases — hyphens, dashes, numbers — are handled slightly differently. Core counts agree; only edge cases vary by a word or two.
How is Chinese or Japanese counted?
By characters, not words — CJK languages do not use spaces, so character count is the meaningful metric.