Word & Character Counter
Words, characters, sentences, paragraphs and reading time as you type — with live social-media character limits and a word-frequency table (stop-word toggle + CSV export). Nothing you paste is uploaded.
Live counts
Social & SEO character limits
Word frequency & text analysis
Paste text above to see which words you use most, each with a proportional bar.
| Word | Count | Share |
|---|
Words are case-folded (Don't and don't are one word) and surrounding punctuation is stripped, but an apostrophe or hyphen inside a word is kept, so don't and e-mail stay whole; letters in any script are counted (café, naïve, 日本語). Share is each word's fraction of all words. The stop-word toggle hides a short, editable common-English list (the, and, of, a, to, in, is, it…) so your real topic words stand out — it is a convenience list, not an authoritative one. Reading time is words ÷ 200 wpm.
Nothing leaves your browser. Your text is analysed on your own device with plain JavaScript. There are no network requests, no logging, and no analytics on this tool — you can disconnect from the internet and it still works.
Who this word counter is for
Writers, students, marketers and developers keep hitting the same walls: an essay with a hard 1,500-word ceiling, a tweet that must land inside 280 characters, a meta description Google will quietly truncate, or a blog post where one keyword has taken over the page. This tool answers all of those at once, live, without sending your draft anywhere. Paste your text and the words, characters, sentences, paragraphs, reading time and keyword mix update on every keystroke — no upload, no Analyse button to hunt for.
Worked example: reading time for a 1,500-word article
Reading time is just word count divided by a reading speed, rounded up to the next whole minute. This tool uses 200 words per minute (wpm), a deliberately conservative figure for silent reading of ordinary prose. So a 1,500-word article works out to 1500 ÷ 200 = 7.5, which rounds up to 8 minutes. Adult silent-reading speed for general text typically falls between 200 and 250 wpm, so the honest range for the same article is:
| Reading speed | 1,500 words | Who reads this fast |
|---|---|---|
| 200 wpm | 7.5 min (this tool) | Careful / non-native reader |
| 238 wpm | 6.3 min | Average adult, general prose |
| 300 wpm | 5.0 min | Fast, skimming familiar material |
The tool picks the low end on purpose: an estimate that says "8 minutes" and takes a reader 6 feels generous, whereas one that promises 5 and takes 8 feels like a lie. Treat it as a friendly signpost, not a stopwatch — technical or dense material with equations, code or unfamiliar names always runs slower than the raw word count suggests.
How each count is defined
Small definition differences are exactly why two "word counters" disagree by a few percent on the same paragraph. Here is precisely what this one does, so the numbers are reproducible:
| Unit | Definition here |
|---|---|
| Words | Any run of non-space characters. Text is split on whitespace, and non-empty pieces are counted, so don't, e-mail and sign-up each count as one word. |
| Characters | Exact length including spaces and line breaks — the figure most social platforms count against. |
| Chars (no spaces) | The same total with all whitespace removed; useful for design and typography work. |
| Sentences | Counted by terminating ., ! or ?. A run of them ends one sentence. |
| Paragraphs | Blocks separated by one or more blank lines. |
| Avg word length | Total word characters ÷ word count. |
Character limits that trip people up
The limits panel measures your character count (not words) against the caps people overshoot most often: X / Twitter at 280, an Instagram caption at 2,200, a LinkedIn post at 3,000, and a meta description where search engines usually render only the first ~160 characters (and roughly the first 60 of a title). Each bar fills as you approach the cap, turns amber inside the final 10%, and turns red the instant you go over — so you trim before you paste, not after the post is rejected. A subtle catch: X counts every emoji as 2 characters, and a URL as a fixed 23 regardless of its real length, so a post that reads "270" here can still be over on the platform if it is emoji-heavy.
Reading the word-frequency table
The word-frequency table shows how often each word appears relative to everything else you wrote. Each word is case-folded (so Don't and don't are counted together) and surrounding punctuation is stripped, while an apostrophe or hyphen inside a word is kept — don't and e-mail stay whole — and letters in any script are counted, so café and 日本語 are handled. Each row carries a count, a share (that word's fraction of all words), and a proportional bar so the shape of your vocabulary is visible at a glance. In a 1,500-word post, a term showing up 30 times is a 2% share.
Function words like the, and and of normally dominate any text, so toggle Hide common words to drop a short, editable common-English stop-word list and surface the terms that actually carry your topic. That list is a convenience, not an authority — it is a few dozen everyday words, and real editorial judgement still beats any fixed list. Use Download CSV to pull the visible rows (as word,count,share_percent) into a spreadsheet. There is no magic "correct" density figure — modern search engines do not reward a target percentage — but the table is a fast way to catch unintentional repetition ("just" appearing 40 times), confirm an article focuses on its topic, or spot obvious over-stuffing.
Common mistakes
- Assuming every counter agrees. Word processors, CMS editors and this tool split on slightly different rules around hyphens, em-dashes and numbers. If an assignment says "1,500 words in Microsoft Word", check the final count in Word — treat this as a live working estimate.
- Confusing characters with characters-no-spaces. Twitter counts spaces; a "500 character" design field might not. Read which number a limit actually means.
- Trusting reading time as exact. It is word count ÷ 200, nothing more — it cannot know your reader's speed or how hard your material is.
- Chasing a keyword-density number. Writing to hit "2%" produces stilted prose and no ranking benefit; use the table to spot repetition, not to game it.
Honest limits
Sentence and paragraph counts are punctuation-based heuristics, not grammar. A sentence ending in "etc." or "U.S.A." can be miscounted, and a list without blank lines reads as one paragraph. Reading time assumes uninterrupted silent reading of prose — it is not calibrated for poetry, tables, or code. And the counts are for the text you paste as-is: markdown symbols, HTML tags or template placeholders all count as ordinary characters. None of that data ever leaves the page — the logic is a small JavaScript module you can read, there are no network requests and no analytics on this tool, so closing the tab is enough to forget your draft.