The same passage has several legitimate lengths: a submission form counts characters, a translation agency counts English words, and an editor counts Han characters. This tool reports six of those measures side by side — total characters, characters without whitespace, Chinese characters, English words, lines and paragraphs — recalculating on every keystroke, with no button to press. Everything is measured inside this tab, so the draft you paste is never transmitted.
How to use it
- Type or paste into the large box at the top (its accessible label is "Text input"). There is no "calculate" button — the six figures update as you type.
- Read the six tiles below the box: Total characters, Characters (no spaces), Chinese characters, English words, Lines, and Paragraphs. Each value is thousands-separated, so a long article stays legible.
- Press Clear, at the bottom right of the input box, to empty it before the next draft; all six figures drop back to zero with it.
- Nothing is produced to take away: there is no output box and no copy button, and the figures are gone once you reload the page.
When it comes in handy
Fitting an application or submission limit
A 500-word bio, a 300-character statement of interest, a competition entry capped at 5,000. Which figure applies depends on the form: "Characters (no spaces)" matches most limits, while older systems that charge you for spaces line up with "Total characters". Having both on screen means you never have to guess which convention the form uses.
Hard caps in ads and social copy
A Google Ads headline allows 30 characters, an email subject line gets truncated on phones past a certain length, and a social post collapses behind "See more" after a few lines. Those platforms count one character per letter, which is what "Total characters" reports. Glancing at it as you draft beats writing the whole thing and then hacking it down.
Quoting translation or copywriting work
Agencies and freelance marketplaces almost always price per English word, so "English words" feeds straight into a quote. Note what it actually matches: runs of letters. don't counts as two, ISO 9001 counts as one, and 2024 counts as none — so a spec sheet or price list will come out lower here than in a tool that counts whitespace-delimited tokens.
Checking how many records or paragraphs you have
Paste a spreadsheet column or an exported list and "Lines" gives you the record count faster than scrolling to the bottom for the row number. "Paragraphs" earns its place in long-form writing: if two thousand words report as a single paragraph, you never left a blank line, and the result reads as a wall of text on screen.
Limits and behavior
- Total characters is the JavaScript string length, which counts UTF-16 code units. Ordinary letters, digits, punctuation and common Han characters are one each, but emoji and rare ideographs beyond the Basic Multilingual Plane (
𠀀, for instance) are stored as a surrogate pair and therefore count as two. - Characters (no spaces) strips every whitespace character, not just the ASCII space: tabs, line breaks, the ideographic space
- English words are matched with
\b[a-zA-Z]+\b, which recognises the 26 unaccented letters only.cafétherefore counts as one (the match iscaf),naïvesplits into two, andABC123counts as none at all, because no word boundary forms where a letter runs straight into a digit. - The Chinese-character figure counts Han characters only, across three ranges: CJK Unified Ideographs, Extension A, and Compatibility Ideographs. Hangul and emoji are excluded, and so are Japanese kana — only the kanji in a sentence register. Rare ideographs beyond Extension B (
𠀀and friends) are also left out: they occupy two UTF-16 code units, and including them would mean matching by code point rather than by code unit as the current implementation does. - Lines counts the segments produced by splitting on line breaks, so pressing Enter once more at the end of the text adds one. An entirely empty box reports zero across all six tiles, including lines.
- Paragraphs are separated by at least one blank line, and a run of several blank lines still counts as a single separator. A lone line break does not split a paragraph, so verse or lyrics with Enter after every line report as one paragraph. Content consisting only of whitespace gives zero paragraphs while lines still reflects the blank rows.
- There is no file picker on this page:
.txt,.docxand.pdfall have to be opened elsewhere and copied across. Text pasted out of a PDF usually breaks every forty characters or so, which distorts both the line and paragraph figures — rejoin it with the line break tool first. All counting happens in this tab's JavaScript and your text stays on your device; the page does still load third-party resources such as advertising, and those requests carry nothing from the input box.
Frequently asked questions
Why do these numbers disagree with the word count in Word?
They measure different things. Word reports a single blended figure that treats each Han character and each English word as one unit and folds some punctuation in. Here the two scripts are kept apart in separate tiles, and full-width punctuation such as ，。「」 never enters the Chinese count. When the number you actually have to satisfy comes from a publisher or a submission form, "Characters (no spaces)" is usually the closest match.
I pressed Enter at the end and the line count went up. Is that a bug?
No — the count comes from splitting on the line-break character. a is one line; a followed by a line break splits into a and an empty string, which is two. Subtract one when you paste a list that ends with a trailing newline, and remember that three stray Enters at the end add three lines. That trailing whitespace is the usual reason a record count looks too high.
Why does one emoji count as two characters?
Browsers hold strings as UTF-16, and anything outside the Basic Multilingual Plane — most emoji, plus the rarest extended ideographs — needs two code units, so the string length reports two. It is not a miscount so much as a reminder that "one character" means different things in different systems: a platform counting UTF-8 bytes would score the same emoji as four. When an exact limit matters, paste the finished text into that platform's own field and see.
Can I count just the part I selected?
No. The statistics always describe the entire contents of the input box, and there is no "selection only" option on screen. To measure one passage, press Clear and paste just that passage in — since counting is live, going back and forth costs nothing.
Does my draft get uploaded anywhere?
It does not. All six figures are computed by JavaScript already running in this tab, directly on your device, and no request carries the contents of the input box out. Disconnect from the network and the counting still works, which you are welcome to verify. For completeness: like every page on this site it loads third-party resources such as advertising, but those requests carry page and ad-slot information, not the text you are measuring.