Word & Character Counter
Five live counts for your own text. Check words, characters, whitespace-free characters and sentences, plus a rough token estimate, without sending your text anywhere.
Counts update as you type, in this tab. No text is sent or saved. The limit counts UTF-16 units; an astral emoji uses two. Long pastes are never cut.
Enter text to see live counts. All five counts are zero at empty input.
- Words
- 0
- Characters
- 0
- Characters without spaces
- 0
- Sentences
- 0
- Estimated tokens
- 0
Estimated tokens = ceil(Unicode code points / 4). A rough, model-independent estimate; language and content can change actual token use substantially.
What these numbers mean
- Words
- Runs of non-whitespace, using the same definition as the GenPolish editor. Apostrophes and punctuation do not split a run; CJK text without whitespace counts as one word.
- Characters
- Unicode code points, not visible symbols or UTF-16 units. Combining accents and joined emoji can contain multiple code points. The 20,000-unit input limit is a separate technical measure.
- Characters without spaces
- Code points after excluding ECMAScript whitespace (\s), including ordinary spaces, tabs, line breaks and non-breaking spaces. This is not just a count without ASCII spaces.
- Sentences
- The editor’s heuristic: a period, exclamation mark or question mark followed by an ordinary space, CR, LF or the end of the text. Common abbreviations such as Dr., e.g. and vs. are protected, and a final non-empty fragment counts. A tab or non-breaking space alone does not end a sentence. This is not universal linguistic segmentation.
- Estimated tokens
- ceil(Unicode code points / 4), rounded up. This local approximation is not a tokenizer or tied to a model. Actual counts depend on the language and content.
Worked example
A short status update and the five numbers this page gives it. Paste it into the counter above and you will get the same results.
Our team shipped version 2.4 on Monday. It fixes the export bug reported by Dr. Okafor, e.g. missing page numbers in PDFs. Can you check the release notes before Friday? Thanks!
- Words
- 31
- Each run of non-whitespace is one word, so 2.4, Dr. and e.g. count once each, and punctuation stays with the word it touches.
- Characters
- 177
- Every letter, digit, space and punctuation mark counts once. With no accents or emoji in this paragraph, that matches what you see.
- Characters without spaces
- 147
- 177 minus the 30 spaces between the words.
- Sentences
- 4
- The periods in 2.4, Dr. and e.g. do not end a sentence: no space follows the first, and the other two are protected abbreviations.
- Estimated tokens
- 45
- 177 divided by 4 is 44.25, rounded up to 45.
Edge cases, counted
| Input | Words | Characters | Sentences |
|---|---|---|---|
| Dr. Smith arrived. Next fragmentDr. is a protected abbreviation, and the unfinished last fragment still counts as a sentence. | Words5 | Characters32 | Sentences2 |
| A vs. B? Yes. No!vs. does not end a sentence; the question mark, the period after Yes and the exclamation mark do. | Words5 | Characters17 | Sentences3 |
| 你好世界No whitespace between the four characters, so they count as one word. | Words1 | Characters4 | Sentences1 |
| 👩💻One emoji on screen, three code points: woman, zero-width joiner, laptop. | Words1 | Characters3 | Sentences1 |
| éAn e followed by a combining accent. It displays as é but counts as two characters; a precomposed é counts as one. | Words1 | Characters2 | Sentences1 |
The last three rows are where counters most often disagree. A tool that counts what you see would report one character for the emoji and one for the accented e, and a tool with language-aware word breaks would split the Chinese phrase into words. Neither result is wrong; each answers a different question. Compare figures from the same tool: this page always gives the same five numbers for the same text.
Frequently asked questions
How are words and sentences counted?
A word is a consecutive run of non-whitespace characters, so punctuation on its own can count as a word. Sentences split after a period, exclamation mark or question mark followed by a space, line break or the end of the text, with some common abbreviations protected. A remaining non-empty fragment also counts as a sentence.
How are Unicode characters and spaces handled?
Characters are Unicode code points, not UTF-16 units or visible symbols. A combined emoji or a letter with a combining accent can contain several code points. Characters without spaces excludes JavaScript whitespace, including tabs and line breaks; it does not remove every invisible character.
Why might another counter or tokenizer show a different result?
Other tools may use language-aware word boundaries, visible-character clusters or different sentence rules. This tool uses whitespace-based words, so languages without spaces can produce very different totals. Estimated tokens is the character count divided by four, rounded up, not an actual model tokenizer or a billing calculation. GenPolish billing counts differently: a word longer than 12 characters counts as several, so do not read this page as a price estimate.
Related tools
Want to polish a draft?
Open the editor to review writing suggestions. Your text is not transferred.