Readability Score Calculator
Paste English text to get its Flesch reading ease, grade level, Gunning Fog and SMOG scores, plus word, sentence and syllable counts and reading time. Runs entirely in your browser.
About this tool
Readability formulas answer a narrow, useful question: how much schooling does this text assume? They do it by measuring two things — how long the sentences are, and how many syllables the words have. Neither of those is a measure of quality. A short, punchy paragraph can be empty, and a dense academic sentence can be the clearest possible way to say something complicated. What the scores give you is a warning when the gap between what you meant and what the reader can absorb has grown wider than you thought.
The Flesch reading ease score runs from 0 to 100, higher being easier, and is the one to watch when writing for a general audience. Roughly 60 to 70 is plain English; below 50 the text is working against the reader. The Flesch–Kincaid grade level converts the same two measurements into a US school year, which is easier to act on: if a page aimed at the public scores at grade 14, the sentences are too long, the words too polysyllabic, or both. Gunning Fog and SMOG apply slightly different weights and count "complex" words — three or more syllables — which makes them harsher on jargon.
Both of the underlying measurements are approximations, and it is worth knowing where they break. Sentence splitting counts a full stop followed by a space, so "Dr. Smith went home" reads as two sentences. Syllable counting has no reliable algorithm in English at all: the letters "ea" are one syllable in "bread", two in "create", and the same pattern recurs throughout the language. This tool uses the heuristic common to other readability tools, plus a short list of corrections for frequent words and for suffixes that revive a silent vowel — hopeless should be two syllables, not three. The result is close enough to compare two drafts, and not precise enough to certify a reading age.
The counts below the score are there so you can see the working. If a score looks wrong, the word and syllable totals usually show why, and you can judge whether the formula or its measurement of your text is at fault. The reading times are plain arithmetic: 200 words per minute silently, 130 aloud, which are reasonable midpoints for adult English.
Why SMOG needs at least 30 sentences
SMOG is calibrated on samples of thirty sentences or more, so below that the tool reports "needs 30+ sentences" instead of a number. The other three scores are unaffected and still shown.
The index counts how often a passage contains words of three or more syllables, then scales that count by the square root of the sentence count. That scaling is only meaningful when the sentence count is high enough for a stable rate; on five sentences, a single complex word swings the result wildly.
Rather than print a figure that pretends to be precise, the tool withholds it and says why. The fallback is deliberate: a number that is quietly unreliable is worse than a visible "not enough text", because nobody can tell the two apart by looking.
If you want a SMOG value for a short passage, the honest route is to extend the sample to thirty sentences of similar style — which is exactly what the original method prescribes — rather than to read the number off a short extract.
Long words are counted but never scored
The tool counts words of seven or more letters for information only. None of the four scores use that count; Gunning Fog and SMOG treat a word as complex by its syllables, not by its length.
It is easy to assume a "long words" figure feeds the difficulty score, since length and difficulty look related. But the formulas are built on syllables, and the two do not track: "strengths" is long and one syllable, "idea" is short and three. Measuring difficulty by letters would disagree with every published version of the formulas.
So the long-word count is shown beside the complex-word count as a second view of the same text. Where the two diverge, it usually marks a passage full of consonant clusters rather than one full of jargon — a different writing problem with a different fix.
Why Chinese, Japanese and Korean text is refused
The Flesch family of formulas is calibrated on English, so pasting Chinese, Japanese or Korean text returns an error that explains this, rather than a meaningless score.
Those languages do not separate words with spaces, so extracting "words" from letters yields almost nothing useful. The kind of syllable counting the formulas assume does not apply either. The result would be a number of the right shape and no meaning at all.
The tool separates this case from the ordinary one. A text with no words at all gets one message; text that is recognisably Chinese, Japanese or Korean gets a different one that names the reason. Saying "the formulas are for English" is more useful than "no words found", which sounds like the tool is broken.
Mixed text is treated differently again: if a passage contains any Latin letters as well, it is scored on whatever English words it holds, since a partial answer about the English part is more useful than none.
Syllable counts are estimated, and every tool estimates differently
Flesch-style formulas depend entirely on syllable counts, and English syllables have no reliable letter rule. This implementation checks an exception table first and then applies suffix rules: every counts 2, business counts 2, queue counts 1.
Syllables are not determined by letter count: queue has five vowel letters and one syllable, every has three vowel letters and two syllables rather than three, create is two while created is three. Any rule-based estimate therefore needs an exception table underneath it.
The order here is: the exception table wins, even over the rules, and then suffix rules apply — less, ness, ment, ful and ly each add one, provided the remaining stem still has at least three letters, or words like ally and belly get cut wrongly.
Why it matters: the Flesch reading ease score is 206.835 minus 1.015 times average sentence length minus 84.6 times average syllables per word, and the grade level is 0.39 times sentence length plus 11.8 times syllables per word minus 15.59. One word miscounted by one syllable in a hundred-word text moves average syllables by 0.01, which moves the grade by about 0.12.
That is also why the same article scores differently across tools: not because one is wrong, but because the syllable estimates differ. To compare, use one tool on one set of drafts; to read a trend, do not keep swapping tools.
Frequently asked questions
What is a good Flesch reading ease score?
It depends on the audience. For the general public, 60 to 70 is the usual target — plain English, roughly an eighth or ninth grade level. Anything above 80 reads like a children’s book and may feel simplistic for adults. Below 50 is genuinely hard work, and below 30 means most readers will give up. If you are writing for a specialist audience, a low score is not a fault; the same sentence would be shorter only by dropping the precision.
Why do different tools give me different scores?
Almost always because of syllable counting, which has no exact algorithm and varies between implementations, and sometimes because of how sentences are split. Small differences in those two totals move the score by a few points. What matters is not matching another tool, but watching your own score move as you edit — that comparison is reliable precisely because the same measurement is applied to both versions.
Should I write to hit a particular grade level?
Use a target as a diagnostic, not a rule. If a passage scores far above where you intended, the usual causes are long sentences carrying two or three ideas at once, and abstract nouns doing the work that verbs should do. Splitting the sentences and naming the actor fixes most of it. But simplifying past the point where the meaning is intact produces text that reads easily and says nothing useful.
Why is the score negative, or above 100?
Because the formula is linear and unbounded, while the 0–100 range is only a convention. Very short sentences with one-syllable words push the score above 100; dense academic prose pushes it below zero. The band underneath is clamped to the scale, so a negative score is reported as "very difficult" rather than falling off the end.
Is my text uploaded anywhere?
No. Everything is counted in the page, with no request, no storage and no logging. It keeps working with the network disconnected, which is the simplest way to confirm it.
Related tools
Estimate how long a script takes to read aloud at four speaking paces — Chinese counted in characters, English in words, pauses included. And work backwards from a target length.
Work out your typing speed: WPM on the standard 5-characters-per-word rule, net speed after errors, accuracy and a rating. English in words per minute, Chinese in characters per minute.
Spell any number in English — cardinals, ordinals, currency amounts and years — with US or UK wording. Full accuracy up to 36 digits, no rounding.
Courses with credits and grades — letter or percent — become a weighted 4.0-scale GPA, with per-course breakdown and cumulative GPA. Nothing you type leaves your browser.