How to use the sentence counter
Type or paste text into the box above — the counter updates as you type. Sentence detection is done by a hand-tuned parser that recognizes common English abbreviations, initials, decimals, ellipses, URLs, and quoted sentences so that a period inside Dr. Smith, 3.14, or google.com is not treated as a sentence boundary. When the parser is unsure, it prefers to leave the sentence intact rather than split it incorrectly.
The stats explained
- Sentences. The number of detected sentences.
- Words. Whitespace-separated tokens across the full text.
- Characters. Total characters including spaces and punctuation.
- Avg words / sentence. Total words divided by total sentences. Business and news writing typically averages 15–20; academic writing 20–30. Anything above ~35 is worth reviewing.
- Median words / sentence. The middle sentence length after sorting. Less affected by one huge outlier sentence than the average.
- Longest / shortest sentence. The tool highlights the sentence with the most and fewest words and marks them in the breakdown below.
- Avg characters / sentence. Another way to measure sentence bulk that includes punctuation and long words.
Flesch Reading Ease — what the score means
Flesch Reading Ease combines average sentence length and average syllables per word into a single score. Higher = easier to read.
| Score | Interpretation | Typical text |
|---|---|---|
| 90–100 | Very easy | 5th-grade level; short sentences, monosyllabic words. |
| 80–89 | Easy | 6th grade; conversational. |
| 70–79 | Fairly easy | 7th grade; most consumer web writing. |
| 60–69 | Plain English | 8th–9th grade; Reader's Digest, most journalism. |
| 50–59 | Fairly difficult | 10th–12th grade; Time, The Atlantic. |
| 30–49 | Difficult | College; academic writing. |
| 0–29 | Very difficult | Post-graduate; legal, scientific. |
The score is calculated using a lightweight in-browser syllable estimator (vowel-group heuristic with silent-e correction). It is a reasonable approximation for everyday English but is not a substitute for a linguistic dictionary lookup for edge cases.
How sentence detection works
A simple "split on period" rule fails on real writing — every sentence ending like Dr. Smith arrived at 3:15 p.m. would be split three times. The counter uses these rules instead:
- Split on
.,!,?, or their fullwidth versions。！？, followed by whitespace or end of text. - Do not split when the period follows a common English abbreviation such as Mr., Mrs., Dr., Prof., St., Sr., Jr., etc., i.e., e.g., vs., p.m., a.m., U.S., U.K., or an initial like J.F.K.
- Do not split when the period sits between two digits (decimals like 3.14 or version numbers like 4.0).
- Do not split on the periods inside an ellipsis (… or ...).
- Treat one or more blank lines as a hard sentence boundary (toggleable).
The parser errs on the side of keeping sentences whole; you can spot-check any specific split in the per-sentence breakdown.
When you actually need this
- Trimming long-sentence-heavy paragraphs for clarity.
- Meeting academic or publication style requirements (avg 15–20 words per sentence).
- Estimating reading grade level from a real text sample.
- Checking that a summary uses varied sentence lengths (a mix of short and long reads better).
- Locating outlier sentences you should split or edit.
Frequently asked questions
How does the counter detect sentences?
By splitting on end-of-sentence punctuation while skipping abbreviations, initials, decimals, URLs, and ellipses. See How sentence detection works above for the full rule list.
What is a good average sentence length?
News and business writing average 15–20 words per sentence. Consumer web writing is closer to 10–15. Academic writing pushes toward 20–30. Above 35 words is usually a sign the sentence should be split.
What is median words per sentence?
The middle value when all detected sentence lengths are sorted. Unlike the average, the median is not skewed by a single very long sentence, so it is often a truer measure of "typical" sentence length.
Is the Flesch score reliable?
It is a reasonable approximation of reading difficulty for everyday English. It cannot judge coherence, argument quality, or vocabulary appropriateness — only sentence length and syllable density.
Does a long sentence make a run-on?
No. A run-on joins two independent clauses without correct punctuation — length alone is not the test. That said, sentences over about 40 words are worth reviewing for clarity.
Can it handle very long documents?
Yes. The counter analyzes the full text, and the per-sentence breakdown loads 100 sentences at a time so the page stays responsive.
Does the counter work for languages other than English?
Sentence detection works reasonably for most Latin-script languages (Spanish, French, German, Portuguese, Italian). The abbreviation list is English-tuned, so sentence splits inside foreign abbreviations may be missed. Flesch Reading Ease is defined for English only.
Is my text saved?
No. The tool runs entirely in your browser — nothing is uploaded or logged.
Related text tools
Sources & methodology
- Flesch, R. (1948). "A new readability yardstick." Journal of Applied Psychology, 32(3).
- Common English abbreviation list compiled from The Chicago Manual of Style and Oxford Guide to Style.
- Syllable estimator: vowel-group heuristic with silent-e, diphthong, and consonant-le corrections — a widely used lightweight approximation.