At first glance, counting words seems simple: split on spaces. But real text has hyphens, em-dashes, contractions, line breaks, and punctuation that blur the edges of what a "word" is. Different tools count differently, which is why a Word document, Google Docs, and a web word counter can disagree on the same text. This page explains exactly how this tool counts.
The tool splits text on any sequence of whitespace — spaces, tabs, and line breaks — and filters out tokens that consist entirely of punctuation. What's left is the word count. Hyphenated compounds like "well-known" count as one word. A lone em-dash or asterisk doesn't count. Numbers like "2024" count as one word.
This matches the convention used by Microsoft Word and most academic style guides: a word is a non-whitespace, non-punctuation-only token. Contractions ("don't", "it's") count as one word, not two.
The tool shows two character counts: with spaces and without spaces. "With spaces" counts every character in the text including whitespace. "Without spaces" counts only the non-whitespace characters.
Twitter and many publishing platforms use the "with spaces" count. SEO tools typically use "without spaces" for keyword density. Some academic submission systems specify character limits without spaces. The tool shows both so you can match the requirement of whatever system you're writing for.
A sentence boundary is detected at any period, question mark, or exclamation mark followed by a space or line break. This misses some edge cases — abbreviations like "Dr." or "e.g." can trigger false positives — but it's accurate enough for most prose. The tool errs on the side of overcounting sentences rather than undercounting, which keeps the average sentence length estimate conservative.
Paragraphs are counted as runs of non-empty lines separated by at least one blank line. A wall of text with no blank lines is one paragraph. This matches the standard definition: a paragraph ends when you hit return twice.
The estimated reading time is based on an average adult silent reading speed of about 200–250 words per minute. The tool uses 230 wpm and rounds up to the nearest half-minute. This is a rough estimate. Dense academic text is read more slowly; simple listicles are read faster. It's useful as a rough guide for content planning — "this is a 4-minute read" — not as a precision measurement.
The keyword density panel shows the most-frequent words in the text, excluding common stop words (the, a, and, of, to, in, etc.). A word's density is its count divided by the total word count, expressed as a percentage.
This is useful for two things. First, SEO writers use it to make sure their target keyword appears at an appropriate frequency — enough to be picked up by search engines, not so often it reads as stuffed. Second, editors use it to spot overused words. If "very" appears at 2.4% density, the writing probably needs some trimming.
The stop word list filters the obvious noise. Without it, every text would show "the" and "of" at the top, which is never useful. The remaining words are ranked by frequency.
Writers with submission limits use it to stay within word counts. Content marketers use it for SEO copy analysis. Students use it for essay word limits. Developers use it to check the length of generated text or prompts. Translators use it to estimate project size. It works for any plain text, regardless of the language.