How to use
- Type or paste your text into the box.
- All counts update as you type β there is nothing to submit.
- If your text mixes CJK and Latin script, a breakdown appears showing each separately.
- The frequency table at the bottom shows the most common words and characters.
About counting words
Counting words sounds trivial until the text stops being English. A word counter is usually
built on splitting at whitespace, which works fine for Latin scripts and fails completely for
Chinese, Japanese and Korean β those languages do not put spaces between words, so the whole
paragraph is one token. Many online counters still report 1 for a page of Chinese prose.
This tool counts CJK characters individually and Latin text by word, then adds them. That
matches how each tradition counts: Chinese publishers and academic requirements are stated in
characters (εζ°), while English word limits are stated in words. A mixed document gets both
numbers separately as well as the total.
The other quiet failure is character counting. In JavaScript, "π".length is 2, because the
language counts UTF-16 code units rather than characters. A family emoji built from four
people joined by zero-width joiners reports 11. Anything counting characters with .length
will overcount emoji, and the error compounds in social media drafts where the limit matters
most. This tool segments by grapheme cluster, so one visible character counts as one.
Reading time is deliberately a rough estimate. The 200 words-per-minute figure for English
and 300 characters-per-minute for Chinese are averages across a lot of variation β dense
technical prose reads far slower than a familiar narrative. It is useful for choosing between
a five-minute and a fifteen-minute read, not for anything more precise.
Frequently asked questions
How are Chinese characters counted?
Each character counts as one word. Chinese, Japanese and Korean are written without spaces between words, so there is nothing for a whitespace-based counter to split on β a thousand-character essay comes out as one word. Counting by character is the convention Chinese publishing uses, and it is what this tool does.
Why does my count differ from Microsoft Word?
Word applies its own rules for hyphenated compounds, numbers with units, and East Asian text. This tool treats a hyphenated compound like well-known as one word and counts CJK by character. Small differences on the same text are normal; if you have a hard limit to meet, check which tool the person setting the limit uses.
Are emoji counted correctly?
Yes, and this is where naive counters break. JavaScript's string length counts UTF-16 code units, so a single π counts as 2 and a family emoji π¨βπ©βπ§βπ¦ counts as 11. This tool counts what you actually see: both are one character.
How is reading time calculated?
200 words per minute for Latin text and 300 characters per minute for CJK, which are the common estimates for each. Mixed documents add the two. Treat it as a rough figure β actual reading speed varies hugely with difficulty and familiarity.
Is my text sent anywhere?
No. Everything is counted in your browser as you type. Nothing is uploaded, stored, or logged, which matters if you are counting an unpublished manuscript or anything confidential.
Why is the sentence count approximate?
Because a full stop and an abbreviation look identical. "Dr. Smith arrived." contains one sentence but two periods, and separating them reliably needs a dictionary of abbreviations. The count is close enough to be useful and is labelled approximate rather than presented as exact.