With this tool you can run the six most-used readability algorithms at once: Flesch Reading Ease, Flesch-Kincaid, Gunning Fog, SMOG, Coleman-Liau, and ARI. Along with these results you get an average, so you don't have to decide which formula you need to use to assess readability.
What each algorithm tells you
We've got six algorithms on offer. Each one targets a different audience or use case, so they don't always agree with each other. Looking at all of them together gives a good overview rather than picking one.
Flesch Reading Ease
Rudolf Flesch built this in 1948, and it's the only score where higher means easier. 60 is the rough cutoff for general audiences; below 30 is academic material. This is the only tool that gives you a score rather than a grade level.
Flesch-Kincaid Grade Level
The same as Flesch Reading Ease, recast as a US school grade. It was commissioned by the US Navy in 1975 to grade technical manuals. Microsoft Word's built-in score is based on this, and it's the most widely cited measure in publishing.
Gunning Fog Index
Robert Gunning built this in 1952. It penalises long words harder than Flesch-Kincaid does: a single polysyllabic jargon word drags your Fog score down more than it lowers your Flesch grade. Fog's definition of a "complex word" excludes proper nouns, compound words, and words that only reach three syllables through a common ending like -ing or -ed. Regular novel prose lands around 8.
SMOG Index
SMOG stands for "Simple Measure of Gobbledygook", and it was invented by G. Harry McLaughlin in 1969. Designed for healthcare and government materials where comprehension is critical. Most accurate on samples of 30+ sentences. It's the standard in medical writing.
Coleman-Liau Index
Built by Coleman and Liau in 1975 for digital text. Uses character counts instead of syllables, which avoids the errors prone to syllable-counting heuristics. Often more consistent than Flesch-Kincaid on machine-generated or unusual text, such as jargon-heavy fiction.
Automated Readability Index (ARI)
Also character-based. Developed by Smith and Senter (1967) for the US Air Force to grade printed materials. Standard in US government documents.
How to use the scores
Don't optimize for single score. Write naturally first, then check. If multiple scores cluster around grade 12+ but your audience is general readers, the prose might be too dense. These are some common strategies that lower the grade level (and raise Flesch Reading Ease, since the two move in opposite directions):
- Shorten your sentences (avg 14–18 words is the sweet spot for general prose)
- Opt for using more common words (replace polysyllabic Latinate words with shorter Anglo-Saxon equivalents where possible)
- Prefer active voice (cuts word count and syllables both)
- Break long paragraphs (paragraphs don't affect the math directly, but shorter ones encourage shorter sentences)
Conversely, if you're writing literary fiction or technical prose for an expert audience, a "high" score (grade 12+) may be exactly right.
Reading levels by content type
These are some writing categories and their expected grade levels. Note that the text complexity age shows the minimum age for understanding the text instead of the actual preferred age for the category.| Content | Typical Flesch-Kincaid grade | Text complexity age |
|---|---|---|
| Children's picture books | 1–3 | 6–8 |
| Middle grade fiction | 4–6 | 9–11 |
| YA novels | 5–7 | 10–12 |
| News articles (general press) | 6–9 | 11–14 |
| Adult commercial fiction | 5–9 | 10–14 |
| General-audience blog posts | 7–9 | 12–14 |
| Business / trade publications | 10–12 | 15–17 |
| Adult literary fiction | 8–14 (varies wildly) | 13–19 |
| Academic / scientific papers | 14–18 | 19+ |
No single score tells you whether a page or a chapter is working. Run a few different samples from the same manuscript, watch for scores that swing wildly between them, and treat the numbers as a prompt to reread a section, not a verdict on it. The goal isn't to hit a number, it's to match the reading level your audience actually expects.