Reading time calculator

What is a reading time calculator?

It counts the words in a piece of text and divides by a reading speed to estimate how long that text takes to read. Blogs put the answer at the top of a post, course designers use it to budget a session, and speakers use it to find out whether a script fits the slot.

This one runs entirely in your browser. Nothing is uploaded, there is no account, and the text you paste never leaves your machine.

It differs from the others in two ways. The reading speeds come from a meta-analysis rather than from the round numbers everyone repeats, and the answer is presented as a range, because the underlying research says a single number cannot be honest.

How to use the reading time calculator

1. Paste your text. Any length. The estimate updates as you type, along with the word count, the letter count and the average word length.
2. Pick a reading mode. Silent non-fiction at 238 words per minute, silent fiction at 260, or reading aloud at 183. All three come from the same meta-analysis. Typing your own figure into the words-per-minute box switches to Custom.
3. Read the range, not just the middle. For the two silent modes the panel shows the span across ordinary adult readers as well as the central estimate. Leave the word-length adjustment on unless you want to compare against a flat rate.

The numbers everyone repeats are the wrong ones, and they are not the main problem

The figures in general circulation are 200, 250 and 300 words per minute, and they trace back further than anyone checks. The best available answer comes from Marc Brysbaert's 2019 review in the Journal of Memory and Language, which pooled 190 studies covering 18,573 participants. Adults reading English silently average 238 words per minute for non-fiction and 260 for fiction. Reading aloud, across 77 studies and 5,965 participants, averages 183.

Against 238, the folklore is off by −16.0% at 200 wpm and +26.1% at 300. A 10,000-word report is 50 minutes at 200 wpm, 33 minutes at 300, and 42 minutes at the measured rate. It is worth being fair about the third figure, though: 250 wpm is only 5.0% high, so the most common number in circulation is very nearly right. The 300 wpm claim has the clearest pedigree and the weakest support — it traces to a 2016 paper by Rayner and colleagues that illustrated the point with a table of ten skilled readers averaging 308.

The paper also explains why fiction reads faster, and the explanation is the more useful part. It is word length. English non-fiction averages 4.6 letters per word against fiction's 4.2, and the rate scales in exact inverse proportion: 238 × 4.6 ÷ 4.2 predicts 261 words per minute for fiction, against the 260 actually observed. Rearranged, that equation says something simpler than it looks — people read a constant 1,094.8 letters per minute, and words per minute is just that constant divided by however many letters the words happen to contain. This calculator implements the letters version, which is why the average word length is shown as a row of its own.

We measured it before writing any of this. Across twelve public-domain books, the non-fiction sample averages 4.61 letters per word against the 4.6 the paper reports — a control confirming that this tool counts what the study counted. Word length ran from 3.91 letters in Machiavelli's The Prince to 4.84 in The Federalist Papers, so the adjustment moves an estimate by −15.1% at one end and +5.1% at the other. Our fiction sample came out at 4.33 rather than 4.2, which is a real gap: the sample is nineteenth-century, and Moby-Dick is not representative of what most people read.

And then the measurement reframed the page. Text difficulty turns out to be the small term. Across that corpus, word length spans a factor of 1.24. Ordinary adult reading speed, which the same paper puts at 175 to 300 words per minute, spans a factor of 1.71. So a 1,000-word article takes 4.2 minutes at the average rate and anywhere from 3.3 to 5.7 minutes depending on who is reading it. Which constant a calculator picks is a smaller error than the one every calculator makes by printing a single number at all. That is why this page prints two.

Honest limitations

The estimate is a population average and the tool has no way of knowing you. The range is the honest part of the output; the central figure is only the middle of a distribution that you may sit well outside. If you have ever timed yourself against a reading-time label and found it badly wrong, this is why, and the fault was in the label rather than in you.

Speed depends on what you are doing. Brysbaert found no evidence for distinct reading gears that a skilled reader can shift between — the one real distinction is between genuinely reading a text and scanning it for something, which are different activities rather than different speeds. Nothing here estimates skimming, because the research does not support a number for it.

The word-length equation was fitted on English. A follow-up study on Dutch by Brysbaert, Sui, Duyck and Dirix found the same shape, but only after replacing the English baseline of 4.6 letters with Dutch's own 5.1 — so the constant does not transfer between languages unchanged. Applying these rates to a language other than English gives you an estimate borrowed from English-language research, not a measurement of that language, and this page says so on every version of itself.

It counts letters, not syllables, and it knows nothing else about your document. Layout, images, tables, code blocks, footnotes and how hard the ideas are all change how long a page takes and none of them are visible to a word count. A dense methods section and an equal number of words of dialogue are not the same reading job.

Finally, the presets are figures for adults reading in their first language. The same meta-analysis reports lower rates for children, for older adults and for people reading in a second language. If you are estimating for one of those audiences, the words-per-minute box takes whatever figure you have — and when you supply your own number the tool shows no range, because the spread belongs to the study's population and not to yours.

Why is it free?

Counting words and dividing by a rate is arithmetic, and it happens in your browser. There is no server doing work worth charging for, no upload, no account, and nothing to watermark.

If you want the counts on their own rather than a duration, the word counter on this site gives you those, also without uploading anything.