There is no single definition of a word
Counting words means deciding where one word ends and the next begins. For ordinary English prose the answer is obvious — spaces — and every tool agrees to within a word or two. The differences appear at the edges, and a long document has a lot of edges: links, numbers, abbreviations, hyphenated compounds, dashes, and text in languages that do not use spaces at all.
Word processors traditionally count runs of characters between spaces. Online counters, including the one on this site, usually look for runs of letters and digits, which splits some things that a space-based counter would keep together. Neither approach is wrong; they answer slightly different questions. What matters is knowing which one the person enforcing your limit is using.
Where the counts diverge
Here is how the word counter on this site handles the common edge cases, with the reasoning behind each rule. Other tools may count some of these differently, which is exactly where totals start to drift apart.
| Text | Counted as | Why |
|---|---|---|
e-mail | 1 word | a hyphen inside a word joins it |
don't | 1 word | an apostrophe inside a word joins it |
3.14 | 1 word | a decimal point between digits joins them |
word—word | 2 words | an em dash separates words even without spaces |
and/or | 2 words | a slash separates words |
10:30 | 2 words | a colon separates the numbers |
U.S.A. | 3 words | full stops between letters separate them |
[email protected] | 3 words | the @ and the dot separate the parts |
https://example.com/a-b | 4 words | the address is split at : / and . |
A space-based counter would count every row of this table as one word. On an essay of plain prose the difference is negligible, but a technical document full of URLs, file paths and version numbers can come out several percent higher here than in a word processor. If your limit is strict, check in the tool the assessor will use.
Numbers, symbols and punctuation
Numbers count as words almost everywhere. Punctuation on its own — a spaced dash, a bullet, an ellipsis — does not. Symbols attached to words, such as a trailing percent sign or a hashtag, usually disappear into the word they are attached to.
Chinese, Japanese and Korean
Chinese and Japanese are written without spaces between words, so a space-based count would treat a whole Chinese sentence as one "word". The convention used by Microsoft Word and by most publishers is instead to count each Chinese character and each Japanese kana as one unit. Korean uses spaces between words and is counted like English.
Mixed text is therefore counted in two parts and added together: "Hello world 你好世界" is 2 English words plus 4 Chinese characters, 6 in total. Chinese punctuation is not counted as words but is counted as characters, so "你好，世界！" is 4 words and 6 characters. When a Chinese submission system gives a "字数" limit, find out whether it counts punctuation — some do — and whether it means characters with or without spaces.
Reading time works the same way: this site assumes 238 words per minute for English and 300 characters per minute for Chinese, so 1,500 English words come out at about 6.5 minutes and 3,000 Chinese characters at about 10.
Characters are not simpler
Character limits look more objective than word limits, but they hide their own choices.
With or without spaces
"Hello world 你好世界" is 16 characters counting spaces and 14 without. Forms that say "characters" usually mean including spaces; academic and translation work often uses characters without spaces. The word counter shows both.
What counts as one character
An emoji with a skin-tone modifier, such as 👍🏽, is made of two Unicode code points but appears as one symbol. This site counts what you see — one — while a program that counts code units might count four. An accented letter can also be stored as a letter plus a separate accent mark; it still counts as one here.
Social networks and text messages
Platforms apply their own weighting. On X (formerly Twitter), a post can use 280 units, where most Latin characters count as one and Chinese, Japanese and Korean characters count as two, so a post written entirely in Chinese holds about 140 characters; every link counts as 23 units whatever its length. A standard SMS holds 160 characters in the basic GSM alphabet, but as soon as the message contains a character outside that alphabet — any Chinese character or emoji — the limit drops to 70 per segment, and longer messages are split and billed as several.
One sentence, two counts
To see how the rules add up, take a single sentence from a business report:
Our Q3 report (see https://example.com/q3-report) shows a 12.5% rise in e-mail sign-ups — up from 3,400 to 3,825.This site counts 20 words. A counter that splits only on spaces and skips the lone dash counts 17. The whole gap of 3 comes from the link: here it is split into https, example, com and q3-report, four words, while a space-based counter sees one. Everything else agrees — 12.5%, 3,400 and 3,825 are one word each in both, and so are the hyphenated e-mail and sign-ups.
Now scale that up. A 2,000-word essay with no links comes out the same in both tools. A 2,000-word technical report with thirty links in it can differ by roughly three words per link — ninety or more — which is enough to cross a strict limit. If your text is link-heavy and the limit is tight, move links into footnotes or a reference list — which many limits exclude anyway — or check the count in the assessor's tool before you cut anything.
Which count to trust
- Use the assessor's tool. For a university essay, that is usually Microsoft Word or the submission system itself. For a grant application, it is the online form. Check your final draft there.
- Leave a margin. If your count and theirs differ, the difference is almost always under a few percent for prose. Staying 2–3% under a hard limit absorbs it.
- Know what is included. Titles, footnotes, references, captions and tables are included by some counters and excluded by others — Word, for instance, asks whether to include footnotes. Institutions usually say what counts.
- For characters, ask "with spaces?" and, for Chinese, "including punctuation?".
Getting under a limit
When a draft is over the limit, the word counter on this site shows the count live as you edit, and a word goal turns it into a progress bar. The keyword list shows which words you use most, which often points at repetition worth cutting. Three editing moves save the most words without losing content:
- Remove filler phrases: "in order to" becomes "to", "due to the fact that" becomes "because".
- Replace nouns made from verbs: "make a decision" becomes "decide".
- Cut the first paragraph if it only restates the question.
Keep the original version and compare the two with the text compare tool before you submit, to make sure nothing important went missing. If you have to reformat headings on the way, the case converter turns a list of headings into consistent Title Case in one go.
Limits
- The rules in the table describe this site's counter. Other tools may split URLs, abbreviations and dashes differently; always verify against the tool your assessor uses.
- Platform rules change. Social network limits and weighting are set by the platforms and may be revised.
- Sentence counts are estimates. Abbreviations with full stops can make a sentence count too high.
Frequently asked questions
Why does Word show a different count from an online counter?
Word counts runs of characters between spaces. Many online counters, including this one, count runs of letters and digits, which splits URLs, slashes and dashes into several words. Plain prose comes out almost the same.
How are Chinese words counted?
Each Chinese character counts as one, following Microsoft Word and most publishers. Punctuation is not counted as a word.
Does a character limit include spaces?
Usually yes on forms, often no in translation and academic settings. Check the instructions; the word counter shows both numbers.