Word Counter

Count words, characters, sentences, paragraphs

Paste any text. Get character / word / sentence / paragraph counts, reading time, speaking time, and a Flesch reading-ease score. 100% client-side — your text never leaves the browser.

100% client-side · no network · no save
Characters
0
— w/o spaces
Words
0
— unique
Sentences
0
— avg length
Paragraphs
0
— avg words/¶
Reading time
0s
@ 200 wpm
Speaking time
0s
@ 130 wpm
Flesch reading-ease
—
—
Long words (6+)
0
—
Avg syllables
—
per word

Word counts, reading time, and Flesch: how to use them without obsessing

A word counter is one of those tools that every writer, student, and content marketer has open in a tab. The simple version is "is my essay long enough" — the more useful version is "is my writing readable". This essay covers the metrics this page computes, what they actually measure, when they matter, and when they're noise.

Character counts (with and without spaces)

Most platforms that have a character limit (Twitter historically, LinkedIn headlines, SMS, meta descriptions, ad copy) count characters with spaces. Some (Twitter's media pool, GitHub commit titles) count without spaces. The two numbers are usually close. The with-spaces count is the one you almost always want to check first.

A 280-character tweet typically allows for 30-50 words of text. Not a lot. Short writing benefits from compression more than length: cut every word that doesn't add information.

Words and unique words

Word count is the most-used metric. It's also the most-cited incorrectly: "the average novel is 80,000 words" — true for adult fiction; picture books are 500-1500; middle grade is 25,000-40,000; young adult is 55,000-80,000. The genre matters.

Unique-word percentage (unique / total) is a rough proxy for vocabulary diversity. Technical writing: 30-40% unique. Fiction: 50-60%. Poetry: 70%+. Numbers above 60% in non-fiction usually signal that the writer is varying vocabulary for its own sake — sometimes good, sometimes an obstacle. Numbers below 30% in fiction usually signal a word you've leaned on too hard (find it and cut it).

Reading time and speaking time

Reading time at 200 wpm is the average silent reading speed for adult native English readers on typical material. Fiction readers average 250 wpm; technical readers average 180 wpm; dense academic readers average 100 wpm. Children and second-language readers are slower.

Speaking time at 130 wpm is the typical public-speaking rate. A 5-minute talk is 650 words; a 10-minute talk is 1300; a 30-minute talk is 3900. If you're a podcast host, these are the numbers to watch. The 130 wpm default is for an experienced speaker; nervous speakers are closer to 100 wpm; auctioneers are 250 wpm.

Flesch reading-ease

Flesch reading-ease is a 0-100 score, higher being easier. Formula:

206.835 - 1.015 * (words / sentences) - 84.6 * (syllables / words)

The two factors that hurt the score are long sentences and long words. The fix is the same as every writing guide has said for a century: short sentences, plain words. The reading-ease score quantifies what feels obvious after you read enough.

Score ranges (US school-grade equivalent):

When the score doesn't matter

Three situations where the Flesch score is misleading and you should ignore it:

Why this tool is 100% client-side

A word counter that sends your text to a server is a logger. The server sees every paragraph you've ever counted, which is a single point of failure for whatever you've been writing — emails to your lawyer, draft cover letters, half-finished novels. We don't store your text; we don't even have a backend for this page. The HTML and JavaScript load once when you open the page; the count runs locally; you close the tab; the text is gone. Verify by opening DevTools (F12) and watching the network tab — no requests fire after the initial page load.

Same pattern as /random-color and /password: nothing leaves the browser.

Related SiteTrace tools: Random Color, Password Generator, Security Headers Checker.