Word and Character Counter

Why word counts differ between tools

Paste the same text into three counters and you will often get three numbers. The disagreement comes from edge cases, and knowing which convention a tool uses saves confusion when a word count has to hit a target.

Hyphenated words are the main culprit: is state-of-the-art one word or four? This counter treats it as one, matching Microsoft Word and Google Docs. Numbers are counted as words, which is also the Word convention. URLs count as one word each. Em dashes between words without spaces do not join them.

For academic or contractual word limits, the number that matters is whatever the submission system produces — usually Word's. This counter matches Word closely enough for planning, but if you are within a few words of a hard limit, check in the tool that will actually measure it.

Character counts are less ambiguous but still worth watching: the count including spaces is what social platforms and meta description limits use, while the count excluding spaces is what some academic limits specify. Both are shown.

Reading time and readability, honestly

Reading time here uses 238 words per minute, which is the average for adults reading non-technical prose in a well-replicated meta-analysis. Real speed varies enormously — technical material with code samples might read at half that rate, and a skimmer moves at several times it. Treat the figure as a rough signal for the reader, which is all it is ever used for, rather than a measurement.

Readability uses Flesch Reading Ease, which scores from 0 to 100 based on sentence length and syllables per word. Higher is easier. Around 60 to 70 is standard for general audiences; below 30 is heavy going. It is worth being clear about its limits: it counts syllables and sentence lengths, and knows nothing about whether your writing is actually clear. You can score well with meaningless short sentences and badly with excellent technical prose where the long words are the correct ones.

Use it as a prompt to look, not as a target to hit. A sudden drop usually means a paragraph has grown a 60-word sentence, and that is genuinely worth fixing. Rewriting good technical writing to raise a score is not.

Keyword density is a bad target

The density panel exists because people ask for it, and it is worth saying plainly what it is good for.

There is no optimal keyword density. The idea that Google rewards a particular percentage is a relic of search engines from twenty years ago, and modern ranking systems work on semantic understanding rather than term frequency. Writing to hit two per cent produces text that reads badly, and reading badly has real costs that outrank any imagined benefit.

What the panel is genuinely useful for is diagnosis. If your target phrase appears zero times in a 2,000-word article, the page may not be clearly about what you think it is. If it appears forty times, the writing has almost certainly become repetitive and unpleasant. And the two-word phrase list often surfaces something more interesting: the phrases you actually used, which may not be the ones you intended to target.

The better use of this data is to check coverage of related concepts rather than repetition of one term. An article about token counting that never mentions context windows, pricing or tokenizers is thin regardless of how often it says "token".

How to use the Word Counter

1. Paste or type your text

   Everything updates live. Nothing is uploaded, so drafts and confidential documents are safe to paste.

2. Check the counts you need

   Characters with spaces for social and meta limits, without spaces for some academic limits, words for most editorial targets.

3. Look at readability if the text feels heavy

   A low score usually points at one or two very long sentences. Find and split them rather than rewriting everything.

4. Use density as a diagnostic

   Look for terms that are absent when they should be present, or repeated to the point of awkwardness. Do not write toward a percentage.

Frequently asked questions