Text readability

What it does

It works out how readable a text is and shows you where it snags: the over-long sentences, the average words per sentence, the share of long words, and which words you are repeating.

The two scales

In Italian it is GULPEASE, the Lucisano–Piemontese index: 89 + (300 · sentences − 10 · letters) / words. It runs 0 to 100 and is tied to the reader’s level of schooling — above 80 it can be read with a primary certificate, below 40 it is hard even with an upper secondary diploma.

In English it is the Flesch Reading Ease: 206.835 − 1.015 · (words/sentences) − 84.6 · (syllables/words). It also runs 0 to 100, but it is tied to American school grades.

They are two different scales and they do not compare. The same text gets two values from them that do not mean the same thing: GULPEASE counts letters, Flesch counts syllables. That is why the name of the scale is always next to the number.

On the language

You choose it, it is not guessed. Detecting the language from the text goes wrong exactly on short texts, which are the ones people paste here — and getting it wrong means applying the wrong formula and showing a perfectly plausible number. A menu with two entries is not a burden.

On long sentences

Over 25 words a sentence gets marked, over 35 it gets marked more. These are not errors: that is the length past which a reader has to go back to find the subject again.

Each of the two thresholds carries two signs — a tint and an underline, single or double — so they can be told apart without separating colours.

On the full stop that does not stop

“Mr. Smith arrived.” is one sentence, not two. The commonest abbreviations are in a list and their full stop does not count as an ending: counting it would push the index up, meaning the text would look easier than it is.

The list is deliberately short. Every extra entry is a full stop that stops closing a sentence even when it really does.

On English syllables

They are counted with a heuristic, not a phonetic dictionary: read and read are spelled the same, and a real dictionary would be a file of megabytes for a tool that lives in a tab.

On the average across a text it works — single errors cancel out — on a single word it can be wrong. For Italian the question does not arise: GULPEASE counts letters.

On repeated words

Articles, prepositions and conjunctions are left out. Without removing them, the most repeated words of any English text would be “the, of, and, to, a” — always, in every text, which is a list that says nothing about this one.

What is not here

An index is a hint, not a grade: it says how easy a text is to decode, not whether it is well written. A text made entirely of five-word sentences scores very high and can be unreadable for other reasons.

Italian and English only: the two formulas are tuned to those two languages, and using them elsewhere gives a number, not a measurement.

On your data

Everything happens in your browser. The text is not stored, it never goes into the address, and it is not sent: closing the page clears it.