Sentence length analyzer.

How sentence length is measured

Sentences are found with the browser's built-in Intl.Segmenter using granularity: "sentence", and words inside each sentence with the same object at word granularity, counting only word-like segments (punctuation, spaces and symbols are not words). This follows the Unicode text-segmentation rules your browser ships, so abbreviations, decimals and quotes are handled by those rules rather than by a home-made regular expression. A sentence that contains no words is skipped. Lowercase text after a full stop is not treated as a new sentence by those rules, so check unusual formatting.

Statistics reported

• Mean: total words divided by sentence count.
• Median and 90th percentile: computed with the NIST/SEMATECH e-Handbook percentile definition (rank p(N+1), linear interpolation, clamped to the shortest and longest value). The median is the 50th percentile. Percentiles can fall between two whole numbers.
• Histogram: sentences grouped by word count into bins of the width you choose. The bin width is a display choice, not a standard.
• Longest sentences: the five sentences with the most words, with their position in the text.

What this is not

No target length, pass mark or grade is applied. For Flesch scores and grade levels use the readability checker; for characters, paragraphs and reading time use the word count checker.

Sources: ECMA-402, Segmenter Objects; NIST/SEMATECH e-Handbook, Percentiles. For the broader baseline first, count words, characters, sentences and paragraphs with the Word Count Checker, or browse all text analysis and writing tools.