Syllable counting is approximate. The estimator uses regex heuristics (vowel groups, silent-e removal) and gets ~85% accuracy on real text. Names, proper nouns, and technical terms throw it off. Expect ±15% wobble. Sentence detection is naïve. "Mr.", "U.S.", "3.14", and ellipses can inflate the sentence count, which lowers reading difficulty falsely. Tools differ — Microsoft Word, Hemingway, and this one will disagree by a point or two. Code and tables wreck the numbers. Markdown code blocks, lists, and tables have very different structure to prose. Paste only the prose if you want a meaningful score. Non-English text is unreliable. All these formulas were calibrated on English. Spanish, German, Japanese, Hindi etc. have different word/syllable distributions; the same formula can give wildly off numbers. Use as a relative measure within the same language at best. Short text isn't statistically meaningful. Below ~100 words the variance dwarfs the signal. SMOG specifically recommends ≥30 sentences for accuracy. Readability ≠ quality. "See spot run" scores extremely easy and isn't great writing. Aim for the reading level your audience actually has, then optimise for clarity within that band. Scores can be gamed. Replacing every "is" with "constitutes" raises the Fog index; doesn't make your writing better. Use scores as a check, not a target.

Consumer-facing prose, broad audience: Flesch-Kincaid 7–8, Flesch Reading Ease 60–70. Healthcare patient materials (US): Flesch-Kincaid 6 or below (CDC, NIH guidance). Children's reading material: Match the target grade exactly. Grade 3 means Grade 3 — overshooting by even one level matters. Academic / professional: 12–16 is normal. Don't try to drag a law review article down to 8; the audience expects density. Web copy, marketing, news: Aim for Grade 8. Most newspapers (NYT, BBC) land around 9–10.

Readability Checker

Compute Flesch Reading Ease, Flesch-Kincaid Grade Level, Gunning Fog, SMOG, and Coleman-Liau scores on your text. Spot dense paragraphs, target grade-level appropriately.

Text

Basic stats

Characters

Words

Sentences

Syllables (est.)

Complex words (3+ syl)

Paragraphs

Readability scores

Formula	Score	Interpretation
Flesch Reading Ease	—	Higher = easier (0–100). 60–70 = plain English.
Flesch-Kincaid Grade Level	—	US grade level needed to read the text.
Gunning Fog Index	—	Years of formal education needed.
SMOG Index	—	Years of education — designed for healthcare prose.
Coleman-Liau Index	—	US grade level (character-based, no syllables).
Automated Readability Index	—	US grade level (character + sentence based).
Average grade level	—	—

Per-paragraph breakdown (Flesch Reading Ease per paragraph — densest are red)

Honesty note

Readability scores are heuristics — they were designed for English prose. Code, technical jargon, lists of short items, and non-English text will produce unreliable numbers. Syllable counting uses a regex-based estimator and is approximately ±15% accurate. Use these scores to spot relative differences and find dense paragraphs, not as exact reading-grade decrees.

What is this for?

A readability score is one number that estimates how much schooling someone needs to read a piece of text. Five different formulas have stuck around — they disagree at the margins, but agreeing on broad strokes is the point. Teachers use them to make sure a passage matches a class's reading level; writers and editors use them to spot dense paragraphs before publishing; accessibility auditors use them to flag content that may exclude readers with cognitive disabilities. Government agencies, the military, and many newspapers have official reading-level targets — Flesch-Kincaid Grade Level 8 is the canonical "plain English" target for U.S. consumer content.

What each score means

Flesch Reading Ease (0–100, higher = easier). 90+ = a 5th-grader can read it; 60–70 = "plain English"; below 30 = academic/legal text. Reader's Digest sits around 65; the Harvard Law Review around 30.
Flesch-Kincaid Grade Level — translates the same formula into a US grade. Grade 8 is the "plain English" target for general audiences.
Gunning Fog Index — years of formal education needed. Heavily weighted on "complex words" (3+ syllables). Often used in healthcare.
SMOG Index — designed for medical/healthcare communications. Counts complex words explicitly. Most accurate when you have ≥30 sentences.
Coleman-Liau Index — uses character counts instead of syllables, so it's more robust on text where syllable counting is wrong (acronyms, names, non-English words mixed in).
Automated Readability Index (ARI) — same intuition as Coleman-Liau; uses characters and sentence length, no syllables.
Average grade level — the mean of the grade-based scores, plus a rough audience label.

How to use the per-paragraph breakdown

Below the scores, each paragraph is shown with its own Flesch Reading Ease score and a coloured stripe down the left: green = easy, yellow = fair, red = dense. This is the most actionable view: read the red ones first. A single dense paragraph in an otherwise readable article is often the place where one sentence has run 40 words long, or where you've used three multi-syllable nouns in a row. Splitting it usually fixes the whole document.

Common gotchas

Syllable counting is approximate. The estimator uses regex heuristics (vowel groups, silent-e removal) and gets ~85% accuracy on real text. Names, proper nouns, and technical terms throw it off. Expect ±15% wobble.
Sentence detection is naïve. "Mr.", "U.S.", "3.14", and ellipses can inflate the sentence count, which lowers reading difficulty falsely. Tools differ — Microsoft Word, Hemingway, and this one will disagree by a point or two.
Code and tables wreck the numbers. Markdown code blocks, lists, and tables have very different structure to prose. Paste only the prose if you want a meaningful score.
Non-English text is unreliable. All these formulas were calibrated on English. Spanish, German, Japanese, Hindi etc. have different word/syllable distributions; the same formula can give wildly off numbers. Use as a relative measure within the same language at best.
Short text isn't statistically meaningful. Below ~100 words the variance dwarfs the signal. SMOG specifically recommends ≥30 sentences for accuracy.
Readability ≠ quality. "See spot run" scores extremely easy and isn't great writing. Aim for the reading level your audience actually has, then optimise for clarity within that band.
Scores can be gamed. Replacing every "is" with "constitutes" raises the Fog index; doesn't make your writing better. Use scores as a check, not a target.

Practical targets

Consumer-facing prose, broad audience: Flesch-Kincaid 7–8, Flesch Reading Ease 60–70.
Healthcare patient materials (US): Flesch-Kincaid 6 or below (CDC, NIH guidance).
Children's reading material: Match the target grade exactly. Grade 3 means Grade 3 — overshooting by even one level matters.
Academic / professional: 12–16 is normal. Don't try to drag a law review article down to 8; the audience expects density.
Web copy, marketing, news: Aim for Grade 8. Most newspapers (NYT, BBC) land around 9–10.