Word Counter & Text Statistics
Count words, characters, sentences, paragraphs and lines, with reading and speaking time estimates and a check against common character limits. All processing happens locally in your browser.
Count words, characters, sentences and reading time
Counting rules are documented below the tool, including how hyphenated words, decimals and abbreviations are handled, so the numbers are reproducible rather than a black box.
What is counted, and exactly how
Word counts differ between tools, and the disagreements are not bugs — they are different definitions. Microsoft Word, Google Docs, a CMS character limit and a linguist's corpus tool will all give you different numbers for the same paragraph.
So rather than presenting a number and asking you to trust it, here are the rules this page uses.
A word
Any run of non-whitespace characters containing at least one letter or digit. Consequences:
| Input | Words | Reason |
|---|---|---|
well-known |
1 | No whitespace inside, so the hyphen does not split it |
state of the art |
4 | Spaces separate, regardless of meaning |
3.5% |
1 | Contains a digit, so it counts |
don't |
1 | The apostrophe is part of the token |
— alone |
0 | No letter or digit present |
e-mail address |
2 | Double spaces collapse; they do not add a word |
A character
Characters are counted by Unicode code point, not by UTF-16 unit. A flag emoji or a rarer CJK character counts as one character here, whereas "🙂".length in JavaScript returns 2. The "without spaces" figure removes every whitespace character including tabs and newlines.
A sentence
A run ending in ., ! or ? followed by whitespace or the end of the text, with a closing quote or bracket allowed after the terminator. Common abbreviations are excluded so they do not inflate the total:
Dr. Patel reviewed the case on Jan. 4. The figure was 3.5% higher.
Naive counting: 4 sentences (splits at Dr. and Jan.)
Counted here: 2 sentencesThe abbreviation list covers titles (Mr., Dr., Prof.), common Latin (e.g., i.e., etc.) and frequent business shortenings (Inc., Ltd., Dept.). It is not exhaustive, so an unusual abbreviation can still add a false sentence.
A paragraph
A block separated by at least one blank line. A single newline is a line break, not a paragraph break — which matches how Markdown, plain-text email and most editors behave. Text with no blank lines at all is one paragraph however long it runs.
How to use the counter
- Paste your text Every figure updates as you type. Nothing needs to be submitted.
- Read the two stat rows The first row covers words, characters, characters without spaces and sentences. The second covers paragraphs, lines, reading time and speaking time.
- Check against a character limit The limits panel compares your character count with common fields — page title, meta description, social post, ad headline — and shows how many characters remain or how far over you are.
- Look at repetition The frequency table lists the most repeated words of four letters or more, with each word's share of the total. It is a drafting aid for spotting a word you have leaned on too heavily.
Reading time and speaking time: what the numbers assume
Both are simple division, and both are estimates rather than measurements.
| Estimate | Rate used | Why that figure |
|---|---|---|
| Reading time | 225 words per minute | Mid-range for silent reading of general prose by adults. Published averages sit roughly between 200 and 260. |
| Speaking time | 140 words per minute | Typical conversational presentation pace. Audiobook narration is nearer 150; a deliberate keynote is closer to 120. |
The formula is words ÷ rate × 60 seconds, with no rounding up to a whole minute, so 90 words shows as 24 sec rather than "1 min read".
What these estimates ignore: reading speed drops sharply for dense technical writing, code samples, tables and unfamiliar terminology — often to half the rate for prose. Time spent on images, diagrams and re-reading is not counted at all. Treat the figure as an order of magnitude, not a promise.
Examples: where the numbers come from
Take this input:
The deploy finished at 14:02. Nothing looked wrong.
Twenty minutes later the error rate climbed.| Measure | Value | Working |
|---|---|---|
| Words | 15 | 14:02 counts as one word because it contains digits |
| Sentences | 3 | Two terminated by . in the first block, one in the second |
| Paragraphs | 2 | Separated by a blank line |
| Lines | 3 | Two text lines plus the empty line between them |
| Average sentence length | 5.0 words | 15 words ÷ 3 sentences |
Note that lines and paragraphs are different numbers for the same text. That distinction is usually what people mean when they say a word counter "gave the wrong answer".
Use cases
- Fitting a character limit. Page titles, meta descriptions, ad headlines and social posts all have hard or effective caps. The limits panel tells you the remaining budget without counting by hand.
- Timing a talk or a script. Speaking time turns a draft into a slot length, which is more useful than word count when you have ten minutes.
- Meeting an editorial brief. Commissioned pieces are specified in words. Average sentence length is a quick proxy for whether the prose has become heavy.
- Trimming repetition in a draft. Seeing that one term accounts for 4% of all words is usually enough to prompt a rewrite.
- Checking transcript length. Comparing word count against known audio duration reveals whether a transcript is complete.
- Sizing a translation job. Rates are typically per source word, so the word count is the quote.
For keyword-focused analysis of a page rather than general text statistics, the Keyword Density Checker is the better fit, and the SERP Preview Tool shows title and description truncation visually rather than numerically.
Privacy
All counting happens in your browser as you type. The text is not uploaded, so unpublished drafts, client work and internal documents stay in the page. Nothing persists after you close the tab.
Frequently Asked Questions
Why does this give a different word count from Microsoft Word?
Word applies its own rules, notably around hyphenated compounds, numbers with separators, footnote markers and text inside shapes or headers. Differences of one or two percent on a long document are normal. What matters for a brief is being consistent about which tool you measure with.
Is a hyphenated word one word or two?
One, here. "Well-known" contains no whitespace, so it is a single token. Some publishers count it as two and some style guides differ, so if you are writing to a strict limit it is worth asking which convention applies.
Why is my emoji counted as one character when my code says two?
Counting here is by Unicode code point. JavaScript strings are UTF-16, so an emoji outside the Basic Multilingual Plane occupies two code units and .length reports 2. Both numbers are correct for their own question; this page answers "how many characters does a reader see".
Does the sentence count handle bullet lists?
Partly. Bullets ending in a full stop are counted as sentences. Bullets with no terminator are counted as one sentence for the whole run, because there is no punctuation to split on. Lists always need a sanity check.
Why does the word frequency table skip short words?
Words under four letters are almost always function words — the, and, for, with — which dominate any frequency count without telling you anything. The threshold keeps the list focused on content words you might actually want to vary.
Can I count the words in a file?
Not directly; this page works on text you paste in. For most formats, opening the file and copying the text across takes a moment and avoids any question about how embedded objects, footnotes or headers should be counted.