Word & Character Counter

How to use it

Type or paste text into the box. Every number updates as you type, and the platform table shows how many characters you have left for each destination. For very long text the update waits a moment after you stop typing so the page stays responsive. Nothing is uploaded: the counting happens in your browser, and the text never leaves your device.

What each number means and how it is counted

Words are counted with the browser's Unicode-aware word segmenter (Intl.Segmenter), keeping only segments that are word-like. That means punctuation alone is not a word, "don't" is one word, numbers count, and Chinese or Japanese text is split into words instead of being treated as a single block. A hyphenated term such as "well-known" counts as two words. Other tools may count differently, so small differences of a few words are normal.

Characters are grapheme clusters, the units a reader sees as one character. A family emoji built from several code points, a flag, or a letter with a combining accent each count as one. "Without spaces" ignores spaces, tabs and line breaks. Sentences come from the same segmenter and ignore empty pieces. Paragraphs are blocks of text separated by at least one blank line; a single line break does not start a new paragraph.

Reading time assumes about 238 words per minute, the average for silent reading of non-fiction in the 2019 meta-analysis by Marc Brysbaert (Journal of Memory and Language). Speaking time assumes about 140 words per minute. That figure is a rough average for presentations and varies widely between speakers, so treat it as a planning estimate.

Platform limits

• X post, 280: X uses a weighted count. This tool approximates the open-source twitter-text configuration: most characters weigh 1, while CJK characters and many other scripts weigh 2, a whole emoji (even a flag, a skin tone or a family made of several code points) weighs 2, and any link counts as 23 no matter how long it is. Only links starting with http(s):// or www. are recognised; bare domains are counted letter by letter. Treat the result as a close estimate and confirm in the composer if you are near the limit.
• LinkedIn post, 3,000 and Instagram caption, 2,200 come from LinkedIn's help center and from Meta's Instagram API documentation (the Instagram app's own help pages don't state a number), as of the review date. They are counted here as characters.
• SEO title (about 60) and meta description (about 160) are display guidelines, not hard limits. Search engines truncate by pixel width, so a title can be cut earlier or later than these numbers suggest.
• SMS: if every character is in the GSM 03.38 alphabet (basic set plus extension set), the message uses GSM-7: 160 characters in a single segment, 153 per segment when split. Extension characters such as € { } [ ] \ ^ ~ | cost two. One character outside that set, such as an emoji or a curly quote, switches the whole message to UCS-2: 70 characters in one segment, 67 per segment when split. Emoji cost two units each in UCS-2.

A few examples

• The text "Meet at café ☕" has 3 words and 14 characters, and in an SMS it forces UCS-2 because of the emoji.
• A 161-character plain English message is two SMS segments, because once it is split each segment holds only 153.
• A post with one 90-character link and 200 other characters counts 223 on X, not 290.
• A 1,000-word article reads in about 4 minutes 12 seconds and takes about 7 minutes 9 seconds to say aloud.

Limitations

Word and sentence boundaries come from your browser's built-in rules, so they can differ slightly between browsers and from word processors. Sentence detection can be fooled by abbreviations. Platform rules change; the X weighting in particular is an approximation, and carriers may segment messages in other ways, for example when a message is sent through a gateway that transliterates characters. Reading and speaking times are averages and do not account for difficult material, technical terms or pauses.