Readability score
Readability score. Runs entirely on your device, nothing is uploaded.
Runs on your device. The file is never uploaded.
Readability score prints Flesch Reading Ease and Flesch-Kincaid Grade Level for a text file to one decimal place, above the word, sentence and syllable counts behind them. A band label sits beside the reading ease figure. Syllables come from a vowel-group rule rather than a dictionary, which is why that count is an estimate.
Questions
Which formula does it use?
Two, both from the Flesch family. Flesch Reading Ease is 206.835 minus 1.015 times words per sentence minus 84.6 times syllables per word. Flesch-Kincaid Grade Level is 0.39 times words per sentence plus 11.8 times syllables per word minus 15.59. Both are reported to one decimal place, along with the word, sentence and estimated syllable counts they were computed from.
How are syllables counted?
By a vowel-group heuristic, not a dictionary. Each word is reduced to its a to z letters, runs of a, e, i, o, u and y are counted as one syllable each, a final e is subtracted unless the word ends in le or the count would drop to zero, and the result is at least one. That gets many words right and is reliably wrong on others, so the syllable figure is labelled an estimate.
What does the reading ease label mean?
It is a band on the Flesch score. 90 and above reads as "very easy", 70 and above "easy", 60 and above "plain English", 50 and above "fairly difficult", 30 and above "difficult", and below 30 "very confusing". Scores are not clamped, so long sentences with long words can produce a negative number, and a file of short simple words can go above 100.
Does it work on languages other than English?
Not meaningfully. Both formulas were calibrated on English, and the syllable heuristic only looks at the letters a to z, so accented text and any non-Latin script contribute zero syllables and skew the result. You will get numbers back for a French or German document, but they do not mean what the labels say.
Why did my score change a lot after minor edits?
Because sentence count drives both formulas and is detected by punctuation. Adding or removing a full stop moves words per sentence, which is multiplied by 1.015 in one formula and 0.39 in the other. Abbreviations, decimal numbers and ellipses all read as sentence ends. If a file has no detectable sentence end at all, the count is treated as 1, so words per sentence becomes the whole word count and the score comes out low.
What happens if my file has no words?
You get "No words found" and a note that there is nothing to score, rather than a division by zero or an invented number. A word here has to start with a letter, so a file holding only digits, punctuation or symbols counts as wordless and is reported that way. Everything runs on your device, and the file is only read, never changed.