Count Words in Text Online

How Word Counting Works

Word counting seems simple at first glance, but the definition of a "word" is surprisingly nuanced. At its core, a word counter splits text on whitespace boundaries and counts the resulting tokens.

The Basic Algorithm

The most common approach uses a regular expression to match sequences of non-whitespace characters:

function countWords(text) {
  const trimmed = text.trim();
  if (trimmed.length === 0) return 0;
  return trimmed.split(/\s+/).length;
}

This splits on one or more whitespace characters (spaces, tabs, newlines) and counts the resulting array length. An empty string check prevents returning 1 for blank input.

Handling Edge Cases

Real-world text introduces several complications:

• Multiple spaces — "hello world" should count as 2 words, not 5. The \s+ regex handles this by matching one or more consecutive whitespace characters.
• Leading/trailing whitespace — Trimming the input before splitting prevents phantom empty strings at the start or end of the array.
• Hyphenated words — Is "well-known" one word or two? Most word counters treat it as one word since the hyphen is not whitespace.
• Contractions — "don't" and "it's" are each counted as a single word because the apostrophe is not a whitespace character.
• Numbers — "There are 42 items" contains 4 words. Numbers surrounded by whitespace are counted as words.

CJK and Non-Latin Scripts

For Chinese, Japanese, and Korean (CJK) text, words are not separated by spaces. A more sophisticated approach uses the Intl.Segmenter API:

const segmenter = new Intl.Segmenter("ja", { granularity: "word" });
const segments = [...segmenter.segment(text)];
const wordCount = segments.filter(s => s.isWordLike).length;

This API uses Unicode-aware segmentation rules that understand word boundaries in non-Latin scripts.

Unicode Word Boundaries

The Unicode standard defines word boundary rules in UAX #29. These rules handle complex cases like emoji sequences (\u{1F468}\u{200D}\u{1F4BB} is one "word"), zero-width joiners, and combining characters. Modern word counters should respect these boundaries for accurate counts across all languages.