Word Counter

Yes. The Chicago Manual of Style, AP, and MLA all count a hyphenated compound (well-being, state-of-the-art) as one word because no whitespace separates the parts. Microsoft Word and Google Docs follow the same rule. The hyphens do not split. For non-hyphenated open compounds (high school), the space genuinely separates two words, so they count as 2.

Yes, in every major style guide. Contractions like don't, isn't, it's, we'd count as one word because no whitespace separates the letters. Microsoft Word, Google Docs, Grammarly, and academic counters all agree. The apostrophe does not split the word - even longer contractions like wouldn't've count as one.

By the whitespace-splitting convention used by Microsoft Word, Google Docs, and this tool, an emoji surrounded by spaces counts as one word. Style guides do not usually address emoji because they predate emoji in formal writing. For character counts, emoji typically take 2 UTF-16 code units, so character-based limits (Twitter 280) charge 2 per emoji.

No. Standard word counters and style guides do not count punctuation as words. "Hello, world!" has 2 words, not 4. Punctuation attached to a word (commas, periods, semicolons) does not split it. Standalone punctuation between spaces would technically split as a token by the whitespace rule, but no style guide treats it as a real word.

It depends on whether the em dash has spaces around it. With spaces around it, the dash separates tokens, so the surrounding text counts as separate words. Without spaces (this-is-a-compound or this—is—joined), the whitespace splitter treats the joined text as one word. Most style guides recommend spaces around em dashes for readability.

For plain English prose with ASCII spaces, yes - all three use whitespace-separated tokens as words and produce the same number on most drafts. They diverge on edge cases: Word counts footnote markers and headings while some online counters do not, and Docs sometimes treats consecutive punctuation differently. For the typical 500-5000 word draft, counts agree within one or two tokens.

The 200 words-per-minute baseline is intentionally conservative. Brysbaert (2019) found adult silent reading averages 238 wpm for non-fiction and 260 for fiction, while academic reading drops to 120-180 wpm. Using 200 keeps the estimate readable for most audiences without underselling dense material. For blog-post "X min read" badges you can safely round; for scripted speech, divide words by 140 instead.

The counter splits on whitespace, and well-formed Chinese, Japanese, and Korean text has no spaces between words. Each character is itself a morpheme, so Chinese editors count characters, not words. The Characters and No Spaces tiles give the meaningful number for CJK. Segmenting CJK into "words" requires a dictionary-based algorithm like Jieba, which a lightweight browser tool does not ship.

The Characters tile uses JavaScript's <code>String.length</code>, which returns UTF-16 code units. Emoji outside the Basic Multilingual Plane - almost every modern emoji, including faces, flags, and ZWJ sequences - occupy two code units, so the counter reports 2 per emoji. A country flag like the French flag is actually two regional indicator emoji joined together, so it counts as 4. Twitter weights emoji similarly when enforcing its 280 limit.

No. The textarea state lives in a Preact <code>useState</code> hook on your own tab. Stats are derived from that state on every keystroke through string operations and regex. Nothing is transmitted to our servers, and the tool does not persist text in localStorage, cookies, or IndexedDB. Closing the tab evaporates the draft; refresh also wipes it unless your browser's form restore kicks in.

SEO studies (HubSpot, Backlinko, Semrush) suggest 1,500-2,500 words for competitive ranking, though reader engagement peaks nearer 1,000. Twitter/X caps free posts at 280 characters; Instagram captions allow 2,200 but favor the first 125 before truncation. Academic essays follow the rubric: a "standard" 2,000-word essay usually allows plus-or-minus 10 percent, and hard-capped theses include appendices separately. Always check the brief; word count conventions vary by publisher.

Partly. The detector splits on <code>.!?</code>, which covers English, Spanish, French, German, and most Latin-script languages. It misses Chinese and Japanese full stops (the ideographic <code>\u3002</code>), Arabic question marks (<code>\u061F</code>), and Armenian punctuation. Abbreviations with periods like "Dr." or "e.g." also inflate the count. Treat the number as a rough guide for English drafts, not a linguistic analysis.

Word auto-corrects straight quotes, apostrophes, and double-hyphens into curly quotes and em dashes. <code>"</code> becomes <code>\u201C</code>, <code>'</code> becomes <code>\u2019</code>, <code>--</code> becomes <code>\u2014</code>. Each smart quote or dash is one character where the original may have been two ASCII characters. Non-breaking spaces from Word (<code>\u00A0</code>) also sneak in and count as characters but not as whitespace splitters.

Usually, but watch for soft hyphens and broken line hyphenation. PDFs often insert a hyphen at line ends that becomes a literal character in your clipboard, so a hyphenated "inform-ation" counts as one long word. Some PDFs also paste with ligatures (fi, fl) as single characters instead of two. Run a find-replace to clean these before trusting the number for a publication target.

Yes. Zero-width joiners (<code>\u200D</code>), zero-width spaces (<code>\u200B</code>), and left-to-right marks (<code>\u200E</code>) all occupy code units, so they bump the character count even though the text looks identical. If your number seems off by a few, paste the content into a hex-dump or inspect-character tool. Translators and localizers run into this constantly.

A paragraph is a block separated from the next by at least one blank line - formally, two consecutive newlines with optional whitespace between them, matched by <code>/\n\s*\n/</code>. Hard line breaks inside a block (Shift+Enter in a rich editor) do not start a new paragraph, and bullet lists where each item is one newline apart collapse into a single paragraph.