Type or paste. Words, characters, sentences, paragraphs, reading time and speaking time update on every keystroke, in this tab, with nothing to click first.
Two tools can read the same paragraph and report different totals, because a word is a definition rather than a fact. Here is the definition this page uses, so you can tell whether a disagreement matters.
A run of letters or digits, with hyphens and apostrophes kept inside it. "State-of-the-art" is one word. "Don’t" is one word, straight apostrophe or curly. Spacing, line breaks and stray punctuation never add to the total.
A full stop, question mark or exclamation mark followed by a space or the end of a block. Common abbreviations are held back, so the full stop after "Dr." or "etc." does not break one sentence into two.
A block of text with a blank line on either side. Single line breaks inside a block stay inside the same paragraph, which is how a pasted document with wrapped lines keeps its real paragraph count.
With spaces counts every character in the box, spaces and line breaks included. Without spaces counts only the rest. Submission forms and character limits use one or the other and rarely say which, so both are always on screen.
One place it will disagree with your word processor: a comma or a decimal point ends a token, so 1,200 and 3.5 each count as two words rather than one. Currency and percent signs are punctuation, so $40 and 40% each count as one.
A word count answers how much you wrote. Reading time answers the question the person on the other end is actually asking, which is how much of their afternoon this is going to take.
Silent reading is calculated at 238 words per minute and reading aloud at 150, which is the pace a rehearsed presenter holds. Both rates are printed beside the figures because they are assumptions about a general reader, not measurements of you or your audience. A dense technical passage runs slower than the first number suggests, and the readability score is the better guide to how much slower.
Type the number into the target field and the bar tracks it as you write. Under the limit it tells you what is left. Over it, it turns red and tells you how much has to come out. Nothing is saved, so the number goes away when the tab does.
Average sentence length sits beside the total. If it climbs while the total falls you are deleting whole short sentences; if it falls, you are tightening the long ones. Those are different edits and they read differently.
A draft that hits the limit can still be one unbroken block. The paragraph and sentence counts sit next to the word count for exactly that reason.
The full report keeps the counts and adds the detectability score, the tells behind it, and inline rewrites you approve one at a time.
Two thousand words can be an easy afternoon or an impassable wall, and the total says nothing about which. The readability checker runs the same text through Flesch Reading Ease, Flesch–Kincaid, Gunning fog and SMOG, shows the sentence-length distribution, and quotes the sentences long enough to lose a reader.
No limit and no account. The counter runs in your browser, so there is no server to meter and nothing to sign up for. Paste a paragraph or a dissertation; the arithmetic is the same.
No. Counting happens in your browser, over a string that stays in your browser. There is no analysis endpoint behind this page, and nothing you paste is sent to a detector.
Because the tools disagree about edge cases, not because one is broken. The main differences are numbers and hyphens: this counter treats a hyphenated compound as one word, and a comma or decimal point inside a number ends the token, so 1,200 counts as two. On ordinary prose the totals match.
Words divided by 238 words per minute for silent reading, and by 150 for reading aloud at a steady presentation pace. Both rates are printed next to the figures. They are assumptions about an average reader, not a measurement of yours.
Yes. Both totals update as you type, and the target field takes a number in words. For a hard character limit, watch the with-spaces figure, which is what almost every form field measures.
Words, characters and paragraphs work for any language written with spaces between words, because the tokeniser matches letters in every alphabet rather than the English ones. Sentence counting looks for a full stop, question mark or exclamation mark, so it holds for languages that use them and not for those that do not. Reading time and syllables are calibrated on English.
Free, unmetered, and everything runs in your browser.