Reading Time Calculator – Turn a Word Count Into Minutes
Almost every publishing platform shows a "5 min read" badge, and almost none of them agree with each other. The same article can be a four minute read on one site and a six minute read on another, because behind the badge there is usually a single division — words / 200 — and nobody tells you which number was used. This reading time calculator keeps that arithmetic but puts every assumption in front of you: the words it counted, the speed it assumed, the time it allowed for images and code, and whether it slowed the estimate down for dense prose.
The core formula
Silent reading time is readingSeconds = (words / wpm) × 60. Read-aloud time uses the same shape with a slower speed, because speaking is a physical act: speakingSeconds = (words / speakingWpm) × 60. A 1,190-word post at 238 words per minute is exactly five minutes of reading, but around eight minutes of speaking at a conversational 150 WPM — which is why a blog post never fits neatly into a video script of the same length.
The 238 WPM default comes from a 2019 meta-analysis of 190 studies of silent reading of English non-fiction, which is a better-grounded figure than the round 200 most platforms assume. It is still an average across a very wide spread of real readers, so the honest way to use any single estimate is as the middle of a range.
Why images and code change the answer
Reading is not the only thing happening on the page. An image stops the eye, and a code block is read line by line rather than swept through. The calculator adds dwell time for both. The Medium model gives the first image 12 seconds and each following image one second less, with a floor of 3 seconds, on the assumption that later images get progressively less attention: imageSeconds = Σ max(12 − i, 3). Code blocks take a flat number of seconds each, 20 by default. Five images and two code blocks add 90 seconds — enough to push a 1,000-word article from a five minute badge to a six minute one.
Complexity: not all words are equal
Long words and long sentences measurably slow readers down. With the complexity adjustment enabled, the tool derives a factor from your average word length awl and average sentence length asl — clamp(1 − 0.03 × max(0, awl − 4.7) − 0.004 × max(0, asl − 17), 0.6, 1) — and applies it to your chosen speed. A legal clause with 6.8-character words and 34-word sentences scores 0.87, dropping an average adult from 238 to about 207 effective WPM, which adds about 38 seconds to every thousand words. The tool also reports a Flesch–Kincaid grade level so you can see the same density expressed as school years.
Word counting, Markdown and CJK
Words are whitespace-separated tokens, so state-of-the-art and don't each count once. Turn off count numbers and bare numerals such as 2026 or 3.14 are excluded. With strip Markdown and HTML on, syntax is removed before counting while the visible text — link labels, image alt text, table cells and the contents of code fences — is kept, so pasting raw source does not inflate the total. Chinese, Japanese and Korean have no spaces between words, so switch the counting mode to characters and the estimate becomes (cjkChars / cpm) × 60.
Writing to a time budget
Reverse mode answers the opposite question. Give it a target — seven minutes, say — and it returns targetMinutes × wpm words, plus a suggested range a few per cent either side, and tracks your draft against it. That is the fastest way to size a conference talk, a newsletter or a video script before you write it. The per-paragraph breakdown then shows which sections are carrying the most time, and the reader comparison table shows the same text as a child, an average adult, a skimmer and a speed reader would experience it. Everything runs in your browser, so unpublished drafts never leave your machine.