Word Counter

Why use a word counter?

Check length limits for papers, posts, listings, and snippets without opening a full editor.

Characters here means UTF-16 code units, so an emoji counts as two

The character count is the length of the text as your browser stores it, and browsers store text as UTF-16. Most characters take one unit, but anything outside the basic range takes two, and some things you see as one character are stored as several. Measured in this tool:

A single 🎉 on its own: 1 word, 2 characters.
café written with the precomposed é: 4 characters.
café written as e plus a combining acute accent: 5 characters.

Those last two are the same word, look identical on screen, and count differently, because the second one is stored as five pieces. Nothing about the text tells you which form you have. If a count is one or two off what you expected and the text contains accents, emoji or non-Latin script, this is why.

A count that matched what a reader sees would have to split the text into what Unicode calls extended grapheme clusters. This tool does not do that, so treat the character count as a storage size rather than as a count of things on the page. For a character limit on a form or a database column, storage size is usually the number you actually want.

A word here is anything between two spaces, so punctuation can be a word

Words are runs of non-whitespace characters. That is simple, predictable, and not the same as what a person would count. Measured:

well-known is 1 word, because the hyphen is not whitespace.
hello,world with no space after the comma is 1 word.
A lone em dash with spaces around it is 1 word, because it is a run of non-whitespace.

So a sentence set with spaced dashes scores higher than the same sentence with unspaced ones, and a missing space after a comma loses you a word. For ordinary prose the total is close enough to be useful; for anything where the number matters, read these rules first.

Non-breaking spaces do count as spaces here. Measured: a and b separated by one non-breaking space is 2 words, 3 characters, 2 characters without spaces. That is worth knowing because our Text Cleaner will not collapse that same character, so the two tools disagree about whether it is space.

A sentence needs punctuation followed by a space, and a quotation mark hides it

Sentences are found by looking for a full stop, question mark or exclamation mark that is either followed by whitespace or sits at the very end of the text. That rule has two failure modes, both measured here, and they pull in opposite directions:

Abbreviations inflate the count. Dr. Smith went home. reports 2 sentences. The full stop after Dr is followed by a space, so it looks like the end of a sentence.
A closing quotation mark hides the end of a sentence. He said "stop." reports 1 sentence, but not because it found one: the full stop is followed by a quotation mark rather than a space or the end of the text, so the pattern matches nothing, and the count falls back to a minimum of 1 for any text that is not empty.

That fallback is why the number never looks obviously broken. Text with no terminal punctuation at all, such as a heading or a list of names, also reports 1 sentence. So read a sentence count of 1 as “one, or none that this tool could see”.

Dialogue and anything with abbreviations, initials, decimals or ellipses will be some distance out. Doing it properly needs sentence-boundary rules per language, which is a much larger job than this tool attempts.

Paragraphs need a blank line, and a trailing newline adds a line

Two different counts here are easy to confuse, so both are worth stating with their measurements:

Two blocks of text separated by a single newline: 1 paragraph, 2 lines.
The same two blocks separated by a blank line: 2 paragraphs, 3 lines.
a followed by a newline: 2 lines, because the newline ends the first line and opens a second, empty one.

So if you write with single line breaks, as people do in a text box, you have one paragraph however many lines you see. And the line count of a file that ends with a newline, which most text files do, will be one higher than the number of lines with anything on them.

The reading time uses 200 words a minute, which is slower than the research says

The estimate here is words divided by 200, shown as <1 min for anything under a minute. We are telling you the number because a reading time with no stated speed is not a measurement, and because our own two tools do not agree.

Measured on the same 1,000 words of text:

This tool: 5 min, at its fixed 200 words a minute.
Our Reading Time Calculator: 4 min 27 sec, at its default of 225 words a minute, which you can change.

The largest review of the evidence puts the real figure higher than either. Brysbaert’s 2019 meta-analysis of 190 studies and 18,573 participants states: “we estimate that the average silent reading rate for adults in English is 238 words per minute (wpm) for non-fiction and 260 wpm for fiction”, and that “for silent reading of English non-fiction most adults fall in the range of 175 to 300 wpm”.

At 238 words a minute those same 1,000 words take about 4 minutes 12 seconds, so the figure on this page is roughly a fifth longer than the average adult would need. Use it as a rough upper bound, and if you need an estimate you can defend, use the calculator above where you can set the speed yourself. Reading rates are lower for children, for older adults and for people reading English as a second language, so no single number fits every audience.

What this tool deliberately does not do

No readability score. There is no Flesch, no grade level and no syllable counting.
No keyword or frequency analysis. For which words repeat, use Word Frequency Counter, which defines a word differently again and will give you a different total.
No grapheme-aware character count, as described above, so the character figure is storage size.
No per-language sentence rules. The sentence count is one regular expression.
No file reading, and no limit. Paste text in; it will keep up with a long document, but it cannot open one.
No saving and no undo. Closing or refreshing the page loses the text.

Nothing is uploaded. Every figure is computed in your browser as you type, and the text never leaves your device.

Sources

Reading rate, 238 wpm for English non-fiction: Brysbaert, M. (2019), “How many words do we read per minute? A review and meta-analysis of reading rate”, Journal of Memory and Language 109. doi.org/10.1016/j.jml.2019.104047 and the full text at reader.ku.edu
Why the character count is UTF-16 code units: String.length. developer.mozilla.org
What counts as one character to a reader, and where word boundaries fall: Unicode Standard Annex #29. unicode.org/reports/tr29
Composed and decomposed spellings of an accented letter: Unicode Standard Annex #15. unicode.org/reports/tr15
The whitespace characters the word and character counts treat as space, including the non-breaking space: ECMAScript, White Space. tc39.es/ecma262
A textarea’s API value and newline normalisation, which is why the line count never sees a carriage return: HTML Standard. html.spec.whatwg.org

Every behaviour and every number on this page was measured against this tool as it is published, in a browser, on 14 October 2026, except the 238 words a minute, which is quoted from the paper cited above. Where the page says what the tool does not do, that is a statement about this tool, not about word counting in general.

Frequently asked questions

Why does one emoji count as two characters?

Because the count is the length of the text in UTF-16 code units, and an emoji needs two of them. Measured: a single 🎉 reports 1 word and 2 characters. The same applies to an accent stored separately from its letter, which is why café can be 4 or 5 characters depending on how it was typed. Treat the character figure as storage size.

Why is “well-known” one word?

Because a word here is a run of characters with no whitespace in it, and a hyphen is not whitespace. By the same rule hello,world with no space after the comma is one word, and a lone em dash with spaces around it is a word of its own. Our Word Frequency Counter splits on hyphens instead, so it counts those differently.

Why is the sentence count wrong?

Two reasons, both measured. Abbreviations inflate it: Dr. Smith went home. reports 2 sentences because the stop after Dr is followed by a space. And a sentence ending inside quotation marks is not seen at all, because the stop is followed by a quote rather than a space, so He said "stop." reports 1 only because any non-empty text reports at least 1.

Why does it say one paragraph when I can see three?

A paragraph break needs a blank line. Blocks separated by a single newline are one paragraph with several lines. Measured: a single newline gives 1 paragraph and 2 lines, a blank line gives 2 paragraphs and 3 lines.

Why is my line count one higher than the number of lines I typed?

Your text ends with a newline, which most text files do. The newline closes the last line and opens an empty one after it, so a followed by a newline counts as 2 lines.

Where does the reading time come from, and can I change it?

It is the word count divided by a fixed 200 words a minute, and no, you cannot change it here. Our Reading Time Calculator lets you set the speed and defaults to 225. On the same 1,000 words this tool says 5 min and that one says 4 min 27 sec. The research average is higher still, at 238 words a minute for non-fiction, so the figure here runs about a fifth long.

Why do your two word tools give different totals?

Because they define a word differently and both definitions are reasonable. On the sentence It's a well-known café — naïve readers don’t notice 2026. this tool reports 10 words and Word Frequency Counter reports 12, because that one splits on accents, hyphens and curly apostrophes. Use this one for a total to quote, and that one for which words repeat.

Is there a length limit, and is my text uploaded?

No limit, and no upload. Everything is counted in your browser as you type, there is no request carrying your text off the page, and nothing is saved, so refreshing loses it.