Live counting as you type
A word counter helps you meet length requirements — tweets (280 chars), meta descriptions (155-160 chars), academic abstracts (250 words), cover letters, essays. Reading and speaking time use 225 words/min (reading) and 140 words/min (speaking), the average for English-speaking adults.
What humans perceive as a single "character" is often internally a sequence of multiple code points. The Unicode standard defines three levels in UAX #29 (Unicode Text Segmentation): bytes, code points and grapheme clusters. A simple letter like A is identical at all three levels (1 byte in ASCII, 1 code point, 1 cluster). A German umlaut like u takes two bytes in UTF-8, is one code point and one cluster. An emoji like (family) is seven code points (four person emojis plus three zero-width joiners) but rendered as a single cluster and thus perceived as one character.
This discrepancy is why different counters produce different numbers for the same text. Microsoft Word counts grapheme clusters ("family" = 1), JavaScript's String.length counts UTF-16 code units ("family" = 11), Python's len() counts code points ("family" = 7). Our tool uses JavaScript's native length, which is correct for plain Latin text but may differ slightly for emojis and CJK. Anyone strictly bound to SMS limits (160 chars GSM-7, 70 chars UCS-2) or Twitter limits (280 code points) should use a grapheme-aware counter.
Word counting has similar ambiguity. Is state-of-the-art one word, three or four? Microsoft Word counts three (each hyphen-separated block), our tool counts one (each whitespace-separated block). Chinese and Japanese lack the "word" concept entirely — the text has no spaces. Our tool returns the number of whitespace blocks, which for Chinese is often 1 or 2 (not meaningful). For Asian text use the character metric instead.
Typical workflow for limit-driven text:
These limits are worth memorizing when writing for the web:
Twitter/X post: 280 characters (Premium: up to 25,000) — images don't count.Meta description: 155-160 chars desktop, 120 mobile — Google truncates beyond that.SEO title tag: 50-60 chars — beyond that a pixel-based cut around ~600 pixels width.LinkedIn post: 3000 characters — but only the first 140-210 are visible in the feed.SMS: 160 chars (GSM-7) or 70 (UCS-2/Unicode) — umlauts and emojis trigger the encoding switch.A counter measures quantity, not quality. 1000 words may be a valuable article or pure filler — the count alone doesn't tell you. The reading time estimate is also an average: 225 words/minute is the median for English adults (source: Brysbaert 2019), but beginners read 100-150 wpm, skilled readers up to 400 wpm, and technical texts (engineering, medicine) often take twice as long. Speaking speed varies even more: news anchors run at 150-180 wpm, conversational at 120-140, podcasts at 130-150. The top-words list carries no semantic meaning either — it just shows frequency. A high count of "you" is normal in marketing text; a high count of a keyword may be helpful or penalized as keyword stuffing. Use the counter as a tool, not as a verdict.
state-of-the-art as three words, we count one. Word counts grapheme clusters as 1 character, we use JavaScript's length. For plain Latin text the differences are minimal.<p> and </p>. To count plain text only, strip HTML first via a browser inspector or a small JS function.?data=... so you can share the link. Nothing reaches our server.