Words, sentences and paragraphs: the exact rules
Word counting sounds trivial until two tools give different numbers for the same text. This counter uses one rule: a word is an unbroken run of letters or digits from any script, and an apostrophe or hyphen inside the run keeps it together. Accented, Cyrillic, Arabic and CJK characters are all letters, although scripts written without spaces count each run between punctuation marks as a single word.
A sentence ends at a period, exclamation mark, question mark or ellipsis followed by whitespace or the end of the text, and only pieces containing a letter or digit are counted. Abbreviations inflate the figure: Dr. Smith is two sentences. A decimal such as 3.5 does not split, because the period is followed by a digit, and text that ends without punctuation still counts as a final sentence.
Paragraphs are separated by a blank line, so a single line break inside a block does not start a new one. Lines are counted literally: the number of line breaks plus one. Pasting from a word processor often gives one paragraph per line with no blank lines between, which shows as one paragraph and many lines. Press Enter twice between blocks if you want them counted separately.
Reading time, word length and the Limits bars
Reading time divides the word count by 238 words per minute and speaking time divides it by 150. Both are flat averages, so a dense technical page takes longer and a list of short lines takes less. Results under a minute show in seconds. 500 words reads as 2 min 6 sec. Average word length counts every character in a word, including an inner hyphen or apostrophe, and Longest word is the first word of maximum length.
The Limits card measures the full character count, spaces included, against five fixed caps: 280 characters for an X post, 160 for a meta description, 60 for an SEO title, 2,200 for an Instagram caption and 3,000 for a LinkedIn post. The bar fills as you type and the figure turns red once you pass the cap.
These caps are plain character counts, exact for some targets and an approximation for others. X weights emoji and many non-Latin characters as two and counts every link at a fixed length, so its composer can reject a post this tool shows as within limit. Search engines truncate titles and descriptions by pixel width rather than by character. The SERP simulator shows how the snippet renders.
Keyword density and what it is good for
The density card groups words case-insensitively, drops anything of one or two letters, drops pure numbers and removes a built-in list of English stop words (articles, pronouns, auxiliaries and common prepositions). The ten most frequent survivors are listed with their count and their percentage of all words. Ties are sorted alphabetically. The card is hidden when nothing qualifies.
Use it as a check rather than a target. If one term sits at 6 percent of a 1,000-word article, it is probably repeated where a pronoun or synonym would read better. Contractions and hyphenated compounds are kept whole, so don't and e-mail appear as themselves, and singular and plural forms are counted separately because the tool does not stem words.
What the transform buttons change
UPPER and lower change the case of the whole text. Title lower-cases everything, then capitalises the first letter of the text and any letter that follows a space, an opening parenthesis, a double quote, an apostrophe or a hyphen. That gives Well-Known for hyphenated compounds and also capitalises the letter after an apostrophe, so check contractions by hand afterwards.
Clean spaces collapses runs of spaces and tabs to a single space, reduces three or more consecutive line breaks to one blank line and trims both ends. Single line breaks are left alone. The .txt button downloads the current text as text.txt, and Clear empties the box. None of the transforms can be undone inside the tool, so keep the original in the online notepad if you may want it back.