Flesch Reading Ease Checker

What the score is for

Reading ease is a crude but useful instrument for one purpose: telling you whether your sentences and your words are long. Editors, plain-language teams and teachers use it as a warning light. It is most useful for comparing two drafts of the same text and least useful as an absolute target: it says nothing about whether writing is clear, accurate, well organised or right for its reader.

The formula

The score comes from the standard formula published by Rudolf Flesch:

Only two quantities matter: average words per sentence and average syllables per word. The coefficients tell you what each lever is worth — removing one word from your average sentence gains about 1 point, while cutting a tenth of a syllable from your average word gains about 8.5 points, so word length dominates. The result is not capped: very short sentences of one-syllable words score above 100, and dense technical prose can go below zero.

The grade labels are a convention from the formula's own literature, not a tested equivalence, and they refer to the US school system. A score near 60 is the usual aim for general-audience writing.

How this page counts, and why that makes it approximate

1. Sentences. The text is cut at every run of ., ! or ? and the non-empty pieces are counted. A full stop at the end of the text therefore does not add an empty sentence — but every abbreviation does add a sentence. Dr., a.m., e.g. and the decimal point in 3.5 each split one sentence into two, which shortens the average sentence and pushes the score up.
2. Words. The text is split on runs of whitespace and the non-empty pieces are counted. Punctuation stays stuck to its word, and a hyphenated compound such as end-of-year is one word.
3. Syllables, step one. Each word is lower-cased. If it is 3 characters or shorter it is given exactly 1 syllable, with no further checks.
4. Syllables, step two. For longer words, one silent ending is stripped if the word ends in any of three patterns: a letter that is not l, y or a vowel followed by es; the letters ed; or that same kind of letter followed by e. This is what stops played and code being counted as two.
5. Syllables, step three. What remains is scanned for runs of the letters a e i o u y, up to two letters at a time, and the number of runs is the syllable count. If no vowel letter is left, the word scores 1.

Counting vowel groups approximates ordinary English well enough to be useful, and is wrong often enough to matter. These are the errors this rule makes:

The errors do not cancel out reliably. Dedicated readability software uses a pronunciation dictionary, so its score for the same text will usually differ by a few points, and by much more for text full of names, acronyms or figures.

Worked examples

The counts and scores below are the values this page produces.

1. Short, plain text. The cat sat on the mat. It was a warm day. 2 sentences, 11 words, 11 estimated syllables. 206.835 − 1.015 × 5.5 − 84.6 × 1.0 = 206.835 − 5.5825 − 84.6, giving 116.65 — above the top of the band table, which is normal for text this simple.
2. Dense official prose. The implementation of the aforementioned regulatory framework necessitates comprehensive documentation. 1 sentence, 10 words, 34 estimated syllables. 206.835 − 1.015 × 10 − 84.6 × 3.4, giving −90.95. Negative scores are possible and simply mean "far harder than the scale was built for". Note that 34 is itself an overcount: "framework" contributed 3 syllables instead of 2.
3. Edge case: abbreviations. Dr. Smith arrived at 9 a.m. He checked the list. A reader sees two sentences; the code finds 4, because Dr. and the two stops in a.m. each end one. With 10 words that gives an average sentence length of 2.5 and a score of 111.24. Spell abbreviations out before scoring, or expect an inflated number.

Practical ways to raise a score

1. Split the longest sentence. Find the sentence with the most commas and make two sentences of it. This is the lever with the most slack in most drafts.
2. Swap long words for short ones. "use" for "utilise", "about" for "approximately", "before" for "prior to", "help" for "facilitate". Because word length carries a coefficient of 84.6, this moves the score fastest.
3. Turn noun phrases back into verbs. "make a determination" becomes "decide"; "provide an explanation" becomes "explain". One verb replaces three or four words.
4. Delete filler and cut the passive where it adds words. "It should be noted that", "in order to" and "due to the fact that" add length without meaning; "the form must be completed by the applicant" becomes "complete the form".

A worked pair: Because the application was submitted after the published deadline, the committee decided that it would be reviewed during the following assessment period. scores 26.84 as one 22-word sentence. Rewritten as The application arrived after the deadline. The committee will review it in the next round. it scores 58.22 — a gain of 31 points from splitting one sentence and shortening six words.

Limits

This works on English only. The syllable rule assumes English spelling and the score bands were built from English text, so running German, Spanish, Hindi or Chinese text through it produces a number with no meaning. Languages that do not separate words with spaces will also count as a single enormous word.

It reports one score and nothing else: no grade level, no Flesch–Kincaid grade, no sentence-by-sentence breakdown, no word or syllable totals, and no suggested rewrite. It does not check spelling, grammar, jargon or factual accuracy — a short, familiar-looking word can still be a term of art. Very short text gives an unstable score, because one long word moves the average sharply; a few hundred words is the least that reads steadily.

Questions

Reviewed October 1, 2026. Calculation methods and corrections