Word Counter

Everything writers count

The limits you're probably writing against

Most counting happens because something enforces a ceiling, so here are the numbers worth keeping nearby: page titles get about 60 characters before search results truncate them and meta descriptions about 155; an X post allows 280 characters; Instagram captions cut their preview after roughly 125; LinkedIn posts fold at about 210 characters until "see more." Academic and editorial work runs on word counts instead — the character panel serves the first group, the word panel the second, and both update on every keystroke so you can edit down against a live target instead of pasting into a form and hoping.

Reading time, speaking time and what they're for

The reading estimate divides your word count by ~225 words per minute, a typical adult prose pace, which is how blogs and newsletters produce their "6 min read" labels. The speaking estimate uses ~130 words per minute — deliberate presentation pace — and is the number to trust when preparing a talk: a 10-minute slot holds roughly 1,300 spoken words, noticeably less than people expect. Rehearsing a script that this page says runs twelve minutes into a ten-minute slot is a cut you can make calmly at your desk rather than on stage.

Reading your keyword density honestly

The top-10 list shows which words carry your text and what share of it each takes. Two habits make it useful: first, check that your actual subject appears among the top words — if you're writing about "invoice automation" and neither word ranks, the piece is probably circling its topic. Second, watch for one word grabbing an outsized share, which reads as repetitive to humans and looks manipulative to search engines. There's no magic percentage; the list is a mirror, and variety plus natural phrasing beats engineering a number.

Step-by-step: editing a draft down to a limit

Paste the draft into the editor; the five stat cards above it — Words, Characters, No spaces, Sentences, Paragraphs — and the header line ("1,240 words · 7,315 characters") fill in immediately. Decide which number your target is measured in: submission forms and social networks count characters with spaces, academic and editorial briefs count words. Keep that card in view and cut from the bottom up — trailing qualifiers, duplicated examples, "in order to" phrasing — watching the number fall after each deletion instead of checking once at the end. Open the Details panel for the secondary signals: reading time, speaking time, line count, average word length and the longest word, which is a quick way to catch an unbroken URL or a hyphenless compound. Finally, glance at Keyword density to confirm the piece still says what it is about after the cuts. Nothing is saved between visits, so copy the final text back out before closing the tab.

Real-world use cases

• Meta descriptions and titles. Type the snippet, watch the Characters card, and stop at about 155 for descriptions and 60 for titles so Google doesn't truncate them mid-word.
• A 10-minute conference talk. Paste the script and read the Speaking time row; at ~130 words per minute, anything past 1,300 words will overrun, and the live count tells you how much to cut before rehearsal.
• University essays with a hard limit. Many institutions allow ±10% of the stated count and exclude references; paste only the body, not the bibliography, to get the number the marker will check.
• Translators and copywriters billed per word. A fast word count on a confidential client document, with the assurance that the text never leaves the machine.
• Newsletter "X min read" labels. The Reading time row is the same 225-words-per-minute convention most publishing platforms use, so your label matches what readers expect.

How the counting works under the hood

A word is any run of non-whitespace characters, so "e-mail" and "3.14" each count once and a stray double space counts nothing — the same rule most word processors apply, which keeps the figure within a percent of theirs. Characters is the raw length of the text including spaces and line breaks; "No spaces" strips every whitespace character. Sentences are counted by scanning for text that ends in a period, exclamation mark, question mark or ellipsis followed by a space or the end of the text, so abbreviations like "e.g." can add a sentence and a paragraph without terminal punctuation still counts as one. Paragraphs are blocks separated by line breaks, ignoring empty lines. Keyword density lowercases the text, keeps only letters, digits and apostrophes, drops words of one or two characters and a list of common stop words (the, and, with, that...), then ranks the remainder; the percentage next to each word is its share of all counted words, not of the raw word total. Reading and speaking times divide the word count by 225 and 130 respectively and round to the nearest second. Because every step runs in a single pass over the text, even a 100,000-word manuscript updates on each keystroke.

Further reading

If the text is headed for a language model, LLM tokens explained shows why token counts differ from word counts; for titles and descriptions that need to fit a preview card, see the Open Graph meta tags guide. Or browse all guides.

Frequently asked questions