Word & Character Counter
Count words, characters, sentences and paragraphs with an estimated reading time, plus the character limits for the places text usually has to fit.
All five rules on your text
The section below shows five defensible definitions disagreeing on a sample paragraph. Here they are on yours — because a spread measured on somebody else's prose is an anecdote, and a spread measured on your own is a decision about which number to quote.
There is no single correct word count
Word count sounds like a fact about a document. It is really a fact about a rule, and the reasonable rules disagree more than you would expect. Here are five defensible definitions run over the same paragraph of ordinary business prose — hyphenated compounds, a few figures, a couple of contractions, nothing unusual.
| Definition | Words | What it does |
|---|---|---|
| Split on whitespace | 42 | What most counters do. Counts a spaced em dash as a word. |
| Runs of letters and digits | 60 | Splits hyphens, commas in numbers and apostrophes. |
| Split hyphens and dashes too | 52 | Treats a hyphenated compound as its parts. |
| Letters, keeping apostrophes | 48 | Contractions stay whole; numbers are not words at all. |
| Unicode word segmentation | 52 | The browser's own rule, via Intl.Segmenter. |
42 to 60 — the highest answer is 43% above the lowest, for one short paragraph. None of the five is wrong. They are answering slightly different questions and none of them says which.
The more useful half of that result is where they agree. Run the same five rules over plain prose — no hyphens, no numbers, no abbreviations, no contractions — and all five return exactly 37. The spread is not vagueness spread evenly through the language. It comes from a small, identifiable set of constructions.
Four constructions cause all of it
Isolating each one shows exactly where the definitions part company. The first row is the control: prose with nothing to argue about, where every rule agrees.
| Text | Whitespace | Letters & digits | Unicode |
|---|---|---|---|
the quick brown fox jumps over the lazy dog today | 10 | 10 | 10 |
a state-of-the-art system | 3 | 6 | 6 |
we spent 3,500 hours | 4 | 5 | 4 |
it cost $12,000.50 exactly | 4 | 6 | 4 |
it isn't ready yet | 4 | 5 | 4 |
Dr. Smith arrived e.g. late | 5 | 6 | 5 |
the plan — such as it was — failed | 9 | 7 | 7 |
Most of those are judgement calls. Whether state-of-the-art is one word or four is
a matter of house style, and whether 3,500 counts as a word at all depends on what
you are counting for.
The last row is not a judgement call — it is a bug. Splitting on whitespace makes a spaced em
dash into a word of its own, so the plan — such as it was — failed comes back as nine
words when it plainly has seven. Any counter whose entire implementation is
text.split(/\s+/) has this defect, and a great many do. This one did too
until we wrote this section; it now requires a letter or digit in anything it counts, which is
why the box above returns seven for that sentence.
What this means for a word limit
A 500-word limit is a measurement plus whichever tool the person checking happens to use. The definitions diverge most on technical and commercial writing — the hyphenated compounds and figures are exactly what that writing is made of — so it is precisely the case where "just under the limit" carries the least information. If a limit is being enforced rather than suggested, it is worth knowing which tool is doing the enforcing.
Reading time inherits the whole problem and adds one of its own, because the words-per-minute figure is an estimate. The same 1,500-word article is an 8-minute read at 200 words per minute and a 5-minute read at 300, and both get printed with equal confidence. The estimate above uses 200 wpm, which is a common figure for careful reading of ordinary prose — treat it as an order of magnitude rather than a promise.
How to use
- Paste or type your text.
- Read the counts as they update live.
- Check character counts with and without spaces.
- Compare against the limit you are writing for.
Frequently asked questions
How is a word counted?
As a run of characters separated by whitespace. That is the standard approach and it makes hyphenated compounds one word, numbers words, and a stray dash a word. Different tools differ slightly at these edges, so two counters disagreeing by a few words on the same text is normal.
How is reading time estimated?
From an average silent reading speed, usually 200 to 250 words per minute for ordinary prose. It is a rough guide only — dense technical writing is read far more slowly, and the figure says nothing about how long the text takes to understand.
Why do character counts differ with and without spaces?
Because different limits count differently. Publishers and academic word limits usually mean characters including spaces; some legacy systems count only visible characters. Where a hard limit applies, it is worth confirming which is meant before trimming.
What are the common limits worth knowing?
A meta description is about 160 characters, a page title about 60, and an SMS 160 characters before it splits. Social platforms change theirs often enough that any figure here would date quickly.
Does an emoji count as one character?
Often not. Many emoji are built from several code points joined together — a family emoji can be seven or more — so a counter may report a single visible symbol as multiple characters. Systems enforcing a limit frequently count the underlying units, which is why an emoji-heavy message hits a limit sooner than it looks like it should.
How many words is a page?
Roughly 250 to 300 double-spaced, or 500 single-spaced, in a standard 12-point font. It varies enough with font, margins and spacing that any word count converted to pages should be treated as an estimate.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.