Local · free · no key
Story word counter
Count a story excerpt with explicit language, title, and Markdown choices. Keep the convention with the number when checking a submission limit.
Explore manuscript preparation →Your input stays in this browser tab. Nothing is uploaded or saved automatically. Export a file before refreshing or leaving if you want to keep your work.
When title is off, the separate title and the first leading # heading are excluded. With Markdown off, common formatting markers, link destinations and code-fence labels are removed; visible code and link labels remain.
Paste an excerpt to begin. Empty text has zero words, sentences and paragraphs.
Calculation rules and limits
Words are Unicode letter/number runs with combining marks; internal apostrophes and hyphens stay within a word. Han characters count individually in Chinese mode; other words use the same rule. Emoji and punctuation are not words. Characters are Unicode code points, so a family emoji may contain several characters; graphemes count visible clusters. Paragraphs are separated by blank lines. Sentence breaks are . ! ? 。！？ or line breaks (an approximate rule, including abbreviations). Positions use 1-based Unicode character offsets and line/column, never byte offsets.
Chinese counting here is not linguistic word segmentation. Sentence splitting is approximate; abbreviations and decimal points may split a sentence. Unicode combining marks count separately as characters but together as visible graphemes. No quality score or reading-grade claim is made.
How to use this tool
- Paste the story and optionally add a separate title.
- Choose language and whether to include the title and Markdown syntax.
- Read the counting rules, then copy or export counts with the selected scope.
Original worked example
Input
One English token 'don't', one Han character 水, and one emoji 🙂.
Output
English mode counts two words: don't and 水. Chinese mode also counts two here, with one Han character; the emoji is not a word. Character and grapheme counts are separate.
Original authored boundary illustration of the published counting convention.
Check your result
- Did you count an added title twice?
- Do language and Markdown settings match the number you intend to report?
Limits and failure cases
- Runs in the browser without uploading text.
- Unicode code points and visible grapheme clusters can differ, especially for combined emoji.
- Counts are operational conventions; a publisher may use another word-count rule.
Continue with your material
Your story context · session only
Keep confirmed facts here, then choose what to use in another task. Refreshing or closing this tab clears this memory. Export a file to keep your work. API keys are never included.