Counting as you type, with nothing uploaded
Paste text in or type into the box and every count updates immediately. There is no button to press and nothing is sent anywhere.
That last part is not a small thing. Most free counters post your text to a server to count it, which is fine for a shopping list and not fine for an unpublished manuscript, a client proposal or a legal letter. This one counts on your device, so the text never leaves the browser.
Why counters disagree with each other
Because "word" and "character" are both less obvious than they look, and different tools make different choices.
The naive way to count words is to split the text on spaces. That over-counts every time there is a double space, a space at the start, or a line break, because each one produces an empty fragment that gets counted as a word. It is why a 500 word essay can come back as 506 in one tool and 500 in another. This one splits on any run of whitespace and then discards anything without a letter or a number in it, so a stray dash between two words is punctuation rather than a word.
Hyphenated words count as one, which matches what a person counting out loud would say. A number counts as a word.
Characters, and the three different answers
A thumbs up emoji is one character to you and two to most software.
That is because computers historically counted characters in UTF-16 units, and anything outside the basic set takes two or more of them. A flag emoji can be four. A family emoji can be eleven. Accented letters are worse, because "café" can be four characters or five depending on whether the é arrived as one character or as an e with a separate accent stuck to it, and the two are indistinguishable on screen.
So this counts what a person means by a character: grapheme clusters, the things you would point at. When that differs from the UTF-16 count, both are shown, because the second number is the one an SMS gateway, a database column or an older form validator will use. If you are writing something with emoji in it that has to fit a limit, the difference matters.
Checking against a limit
The real question behind most visits is not "how many" but "am I under".
So once there is enough text to be near one, the common limits appear with a bar and how much you have left. A page title before Google truncates it, a meta description, a single SMS, an X post, Google Ads headlines and descriptions, an Instagram caption, a LinkedIn headline.
Two caveats worth knowing. Google truncates titles and descriptions by pixel width, not character count, so a title full of capitals and Ms gets cut sooner than the figure suggests: treat it as a guide and keep the important words first. And an SMS with any character outside the basic GSM set drops to 70 characters a message rather than 160, so an emoji in a text can more than halve what fits.
Reading time and speaking time
Reading time uses 238 words a minute, which is the figure from the largest meta-analysis of silent reading speed for English rather than the 200 that gets copied around.
Speaking time uses 130 a minute, which is a measured presentation pace. If you are timing a speech or a video script, that is the number to use, and it is roughly half again as long as reading the same text silently. Both round up to the nearest minute, because "0 min" is not a useful answer for anything with words in it.
Sentences, and where the count is approximate
Sentences are counted on full stops, question marks and exclamation marks, including runs like "?!".
It will over-count when your text contains abbreviations. "e.g." and "Mr." each read as a sentence end. That is a deliberate choice: fixing it needs a dictionary of abbreviations, which then gets a different set of cases wrong instead, and gets them wrong unpredictably. Predictable over-counting is easier to work with than occasional silent errors, and the sentence count is a guide rather than something anybody is held to.
Paragraphs are counted on blank lines rather than every line break, so a block of text with soft wraps is one paragraph. The line count is the editor's, blanks included.
Common questions
Is my text uploaded anywhere to count it?
No. Most free word counters send your text to a server to count it, but this one counts on your device as you type, so nothing is sent anywhere and the counts update with no button to press.
Why do different word counters give different totals for the same text?
Because splitting text on spaces alone over-counts every double space, leading space and line break as an extra word. This one splits on any run of whitespace and discards fragments with no letter or number in them, so a stray dash between two words is not counted as a word of its own.
Why does an emoji count as more than one character elsewhere?
Because most software counts characters in UTF-16 units, and an emoji outside the basic character set can take two or more of them. A thumbs up is one character to a person but two in that system, and a flag emoji can be four. This tool counts what a person would point at, and shows the UTF-16 figure alongside it when the two differ.
Why might an accented word have a different character count?
Because a letter like "café" can be stored as four characters or five, depending on whether the accent arrived as part of the letter or as a separate mark added afterwards. The two look identical on screen but count differently, which is why this tool shows both figures when they diverge.
What reading speed does the reading time use?
238 words a minute, taken from the largest meta-analysis of silent reading speed for English, rather than the commonly quoted 200. Speaking time uses 130 words a minute, a measured presentation pace, which is roughly half again as long as reading the same text silently.
Why does the sentence count go up on text with abbreviations?
Because sentences are counted on full stops, question marks and exclamation marks, and an abbreviation like "e.g." or "Mr." reads as a sentence ending. Fixing that needs a dictionary of abbreviations, which then gets a different set of cases wrong instead, so the count is a guide rather than an exact figure.
Does this check my text against limits like a tweet or a meta description?
Yes. Once there is enough text to be near one, a bar appears showing how much is left against common limits, including a page title, a meta description, an SMS, an X post and a LinkedIn headline. Google truncates titles by pixel width rather than character count, so one full of capital letters can be cut sooner than the number suggests.