How word counts and reading time are actually measured
The exact rules behind word, sentence, and paragraph counts, why tools disagree, and how the 200 wpm reading estimate is built.
What counts as a word
This counter defines a word as any sequence of non-whitespace characters, which is the same rule most editors use. That means numbers, symbols, and hyphenated compounds each register as one word, while a URL with no spaces counts as a single long word. It is a deliberate, predictable rule, but it explains why a technical document full of code snippets or long links can show a lower word count than you expect.
Why two tools can disagree
Word counts vary between apps because each one draws the line differently. Some treat an em dash between words as a separator, some split on every punctuation mark, and others count only dictionary words. Sentence detection is even less standardized, since a period ends a sentence but also marks abbreviations and decimals. Knowing this counter splits strictly on whitespace for words and on runs of . ! ? for sentences lets you predict its output rather than guess.
How the reading time estimate is built
The reading time comes from a simple model: word count divided by 200 words per minute, the accepted average for silent adult reading of everyday prose. Denser material, technical writing, or a second language reads slower, while a skimmer moves faster, so the figure is a midpoint rather than a promise. The result is rounded to one decimal, which is why a very short note can display 0 min even though it takes a moment to read.
Using the counts for real limits
Different platforms enforce different limits, and the right tile matters. Social posts and SMS cap by character, so watch the 'characters' count including spaces. Essays and articles are usually set by word count, and job or school briefs sometimes specify sentences or paragraphs. Because every tile updates live, you can trim toward a target and watch the exact number fall into range as you edit.