Sentence Counter

How to use the sentence counter

This sentence counter does the count a careful reader would do by hand, in an instant and for any length of text: an essay, a cover letter, a report, a speech or a page of a novel. It is also a word counter, a character counter and a readability checker, because a sentence count on its own says little about how a text reads.

1. Type or paste your text into the box, or open a .txt file. The counts update as you type.
2. Read the counts: sentences, words, characters with and without spaces, paragraphs, lines, sentence lengths and reading time.
3. Check the long sentences, marked in your text above the limit you choose, and the six readability scores.
4. Copy the stats or download a CSV report with every sentence and its length.

Press Try a sample on an empty box to see a short paragraph with a title, a time, a decimal and one very long sentence. Under All sentences every sentence is listed with its number and length; select one and it is highlighted in the text box, ready to edit. Counting options holds the reading and speaking speeds and how line breaks are treated.

What counts as a sentence

In everyday writing a sentence ends with a full stop, question mark or exclamation point. G. Harry McLaughlin’s instructions for his SMOG readability formula put it the same way: count as a sentence “any string of words ending with a period, question mark or exclamation point.” The hard part is the full stop, which also ends abbreviations, initials and decimals.

The counter starts from the sentence rules of the Unicode Standard (Annex #29, Unicode Text Segmentation), which your browser implements. Those rules already keep 3.14, example.com and U.S.A together and do not end a sentence before a lowercase word. The Annex itself says they “cannot detect cases such as ‘...Mr. Jones...’”, so the counter adds three corrections of its own:

• Titles and Latin abbreviations never end a sentence: Mr., Mrs., Ms., Dr., Prof., Rev., St., Mt., military and political titles such as Gen., Capt., Gov. and Sen., and e.g., i.e., cf., viz., vs. and approx.
• Initials stay with the name: J. R. R. Tolkien, John F. Kennedy.
• Dialogue tags stay with the quote: in “Stop!” she said, the lowercase “she” shows the sentence goes on.

Each line break also ends a sentence, so a heading or a list item counts as one even without a full stop. For text pasted from a PDF or an e-mail, where lines end in the middle of sentences, choose Ignore single line breaks under Counting options; a blank line still ends the paragraph. These examples were all counted by the tool on this page:

One limit is worth knowing. Because abbreviations like U.S., etc. and Inc. so often end a sentence, the counter ends the sentence when a capital letter follows them, so “The U.S. Army” counts as two. And because a lowercase word after a full stop continues a sentence in the Unicode rules, text typed all in lowercase, such as “i went home. it was late.”, counts as one sentence. Check the All sentences list when a count looks off.

Some style guides and readability instructions also treat a semicolon or a colon as the end of a sentence. The U.S. Navy’s 1975 instructions for the Flesch formula did, and so does Florida’s law on readable insurance policies. This counter does not: a semicolon joins two clauses into one sentence, as most readers see it.

Words, characters, paragraphs and lines

• Words are runs of characters between spaces (or dashes and ellipses) with at least one letter or digit. A contraction, a hyphenated word, an e-mail address or a number such as 32,008 is one word, which is the rule the readability formulas’ own instructions give. A lone dash or ampersand is not a word.
• Characters are counted the way you see them, so an emoji, a flag or an accented letter counts once even when it is stored as several code points. Characters includes spaces and line breaks; characters without spaces leaves out every kind of white space.
• Paragraphs are blocks of text separated by a blank line. Lines are the lines that are not blank.

Character limits are usually counted with spaces, so use the first character count for a bio, a meta description or a form field, and the count without spaces where a brief asks for it explicitly.

Sentence length

Long sentences are the most common reason a text is hard to read. The U.S. Office of Personnel Management’s plain language guidance says “sentence length should average 15-20 words”. The counter shows your average sentence length (words divided by sentences), the longest and shortest sentence, and a chart of how many sentences fall in each length band.

Under the Long sentences tab, every sentence longer than the limit is marked in your text. The limit starts at 25 words, comfortably above the recommended average, so only sentences that really stand out are marked; set it to 20 for a stricter check. An average of 15 to 20 does not mean every sentence should be that long: a mix of short and long sentences reads better than a row of equal ones, and the chart shows that mix at a glance.

Readability scores

A readability formula predicts how hard a text is from two things: how long its sentences are and how long or complex its words are. The counter computes six of the best-known formulas, each exactly as its authors published it. All but Flesch Reading Ease give a U.S. school grade.

Flesch Reading Ease runs the other way: higher is easier, and the score is not clipped, so very simple text can go above 100 and dense text below 0. Flesch described the bands in his 1949 book The Art of Readable Writing:

What each formula counts

• Syllables. Flesch and Flesch–Kincaid use the average number of syllables per word. The counter estimates syllables from spelling (vowel groups, a silent final e, -ed and -es endings, -le, and a list of common exceptions), so no dictionary has to be downloaded. The estimate is right for most common words and can be off by one on names, abbreviations and words people pronounce differently, which is why two checkers seldom agree to the decimal.
• Long words. SMOG counts every word of three or more syllables. Gunning Fog counts the same “hard words” but, following the rules usually given for Gunning’s index, leaves out capitalized names in the middle of a sentence, hyphenated compounds of short words, and words that only reach three syllables through an -ed or -es ending.
• Letters and characters. Coleman–Liau counts letters, and the Automated Readability Index counts characters, which in its original typewriter method included punctuation but not spaces. Neither needs syllables.

The average grade shown above the scores is simply the mean of the five grade-level formulas, a convenience rather than a published formula. Treat every score as an estimate: the authors of the Navy formulas wrote that readability formulas “are only accurate to within one grade level”, and McLaughlin gave SMOG a standard error of about 1.5 grades. SMOG predicts the grade needed for complete understanding, so it usually reads a grade or two above the others. Scores also need enough text: the formulas were built on passages of 100 words or more, and SMOG on 30 sentences, so the tool adds a note below those sizes.

Worked example

Load Try a sample and the counter reads this paragraph:

Dr. Rivera opened the meeting at 9:30 a.m. with one question. How many sentences are in this paragraph? Count them before you read on! She wrote the answer, 3.5 percent, on the board. A careful reader notices that the title, the decimal and the time do not end a sentence, while a hurried reader who stops at every period, without asking what the period is doing there, ends up with a total that is far too high.

It finds 5 sentences and 77 words: the full stops after “Dr”, “a.m” and inside “3.5” do not count. The average sentence is 77 ÷ 5 = 15.4 words, but the last sentence alone has 44 words and is the one marked as long. The text has 109 syllables, 7 words of three or more syllables (Rivera, sentences, paragraph, notices, decimal and period twice) and 4 Fog “hard words” (paragraph, decimal and period twice; Rivera is a name, and sentences and notices only reach three syllables with -es). It has 315 letters and 336 characters in its words.

So the sample reads at about a 7th to 10th grade level, depending on the formula, with an average grade of 7.7. Splitting the last sentence in two would lower every score, which is exactly the kind of edit the long-sentence marks are there to suggest.

Reading and speaking time

Reading time is the word count divided by a reading speed. The defaults come from Marc Brysbaert’s 2019 review of 190 studies with 18,573 participants, which put the average adult silent reading speed for English non-fiction at 238 words per minute (260 for fiction) and reading aloud at 183 words per minute. Most adults read non-fiction at 175 to 300 words per minute, so set your own speed under Counting options if you know it, or a slower speaking speed for a talk with pauses.

Other languages

Sentences, words and characters are found with your browser’s built-in text segmentation (the Intl.Segmenter interface of the ECMAScript internationalization standard), which follows the Unicode rules for every script. It knows the full stops of Chinese and Japanese (。) and the danda of Hindi (।), and finds the words in languages written without spaces, such as Japanese, Chinese and Thai. If a browser does not have it, the counter falls back to its own version of the same rules for sentences.

The abbreviation list and the syllable rules are English, and so are the readability formulas, which were calibrated on English readers. When most of the letters in the text are not Latin, the scores are hidden instead of showing numbers that would mean nothing. Texts in other Latin-script languages are scored, but treat those scores with extra care.

Private by design

• Nothing is uploaded. Every count is made in your browser. The page’s Content-Security-Policy allows connections only back to this site, and the counter makes none; a .txt file you open is read on your device.
• Your text is not stored unless you turn on Save my text in this browser. Then it is kept in this browser’s local storage, under one key, until you turn the switch off or press Forget saved text. Leave it off on a shared computer.
• No account and no cookies. Your settings, such as the long-sentence limit and reading speed, are remembered in the same key. Once the page has loaded, the counter works offline.

Sources

• Unicode Standard Annex #29, Unicode Text Segmentation (Unicode 18.0.0): the default sentence and word boundary rules.
• ECMAScript Internationalization API Specification (ECMA-402): Segmenter objects.
• Flesch R. (1948). A new readability yardstick. Journal of Applied Psychology 32(3), 221–233, with the formula as written into Florida Statutes §627.4145, readable language in insurance policies.
• Kincaid J. P., Fishburne R. P., Rogers R. L., Chissom B. S. (1975). Derivation of new readability formulas for Navy enlisted personnel. Research Branch Report 8-75, Chief of Naval Technical Training: the Flesch–Kincaid grade, the original ARI and the counting instructions.
• McLaughlin G. H. (1969). SMOG Grading: a new readability formula. Journal of Reading 12(8), 639–646.
• Coleman M., Liau T. L. (1975). A computer readability formula designed for machine scoring. Journal of Applied Psychology 60(2), 283–284.
• DuBay W. H. (2004). The Principles of Readability (ERIC ED490073): Gunning’s Fog Index (1952) and Flesch’s reading ease bands (1949).
• Brysbaert M. (2019). How many words do we read per minute? A review and meta-analysis of reading rate. Journal of Memory and Language 109, 104047.
• U.S. Office of Personnel Management: Plain Language.

Reviewed by microplugins · Last updated October 2026