Word Counter

Word Counter is a tool for real-time text statistics including character count, word count, line count, and more. Whether you're writing articles, papers, code comments, or social media content, this tool helps you quickly understand your text. It supports mixed Chinese and English text, intelligently recognizing Chinese characters and English words for separate counting. It also provides reading time estimates to help you understand how long it takes readers to finish your content. A word counter helps with articles, essays, SEO copy, social posts, translations, abstracts, and input limits. Counting rules can differ by language and platform: hyphenated words, numbers, CJK text, emoji, punctuation, and contractions are not always treated the same way. When a requirement is formal, the counting method should be clarified. This tool gives a fast working estimate, but the final count may still need to be checked in the target submission system.

Word counting splits on Unicode whitespace using a regular expression close to `text.trim().split(/\s+/).filter(Boolean)` for Latin scripts, which matches ASCII spaces, tabs, newlines, and Unicode separators like U+00A0 (non-breaking space) and U+2028 (line separator). This rule works for English, French, German, and other space-separated languages, but breaks down for Chinese, Japanese, and Thai where there is no whitespace between words. CJK text is therefore counted by character: `[...text].length` rather than `text.length`, because the spread operator iterates Unicode code points and correctly handles surrogate pairs for characters above U+FFFF (e.g. CJK Unified Ideographs Extension B starting at U+20000 and most emoji at U+1F300+). The CJK character class itself is detected with the Unicode property escape `/\p{Script=Han}/u` for Chinese ideographs. The `length` property of a JavaScript string counts UTF-16 code units, not characters, so `'👨‍👩‍👧'.length === 8` while the visible glyph is one grapheme cluster made up of three emoji joined by U+200D Zero-Width Joiner. The accurate count uses `Intl.Segmenter('en', { granularity: 'grapheme' })`, available in all modern browsers since 2023; for word-level segmentation in any language including CJK, `Intl.Segmenter(locale, { granularity: 'word' })` follows Unicode Standard Annex #29 and is the most correct option when it is available. Sentence boundaries are detected by `[.!?…。！？]+` followed by whitespace or end-of-string, and paragraphs by two or more consecutive line breaks (`/\n\s*\n/`). Reading and speaking estimates apply published reading-rate constants. Brysbaert (2019) meta-analysis puts adult silent reading at 238 words per minute for non-fiction English, which is why 225-250 wpm is the common UI default. Chinese silent reading averages 350-500 characters per minute. Speaking pace is slower: TED talks average 163 wpm, and audiobook narration targets 150-160 wpm. Twitter's 280-character limit counts CJK and most emoji as 2 weighted characters via its `twitter-text` library, while a regional indicator pair (a country-flag emoji) counts as 4 UTF-16 code units in raw `.length`.

• Whitespace word split: `text.trim().split(/\s+/).filter(Boolean)` works for Latin scripts; fails on CJK and Thai which have no inter-word spaces.
• CJK character count uses `[...text].length` to iterate Unicode code points correctly (handles surrogate pairs above U+FFFF, e.g. extended ideographs and emoji).
• `string.length` returns UTF-16 code units, not graphemes: a 👨‍👩‍👧 family emoji has `.length === 8` but is one visible character.
• Most accurate segmentation: `Intl.Segmenter(locale, { granularity: 'word' | 'grapheme' })` implements Unicode Standard Annex #29.
• Reading speed defaults: English 225-250 wpm silent (Brysbaert 2019), Chinese 350-500 chars/min; speaking 150-160 wpm for narration, 163 wpm for TED average.
• Sentence split: `[.!?…。！？]+\s+`; paragraph split: `\n\s*\n`; line count: `text.split('\n').length`.
• Twitter weighs CJK and most emoji as 2 chars against the 280 limit via `twitter-text`; a flag emoji (regional indicator pair) is `.length === 4` in raw UTF-16.