Text counting guide
How word counting works
“How many words is this?” sounds simple until punctuation, hyphens, numbers and unusual formatting enter the text. This guide explains a practical counting method and why two editors can sometimes disagree.
The basic rule
A practical word counter looks for meaningful sequences of letters or numbers separated by whitespace or punctuation. LetterToolkit keeps common internal apostrophes and hyphens inside a word, so don't and well-known are each treated as one word. Extra spaces, tabs and blank lines do not add words.
This is a useful general-purpose rule, not a universal publishing law. A university portal, newsroom or legal filing system may define a word differently.
14 word-count examples you can check
Paste each example separately into the Word Counter. These are the results of LetterToolkit's current counting rule, checked against the tool on October 19, 2026. They are not a claim about the counts shown by other editors.
| Input | Words | Why |
|---|---|---|
| Empty input | 0 | There are no letters or numbers to count. |
Hello, world! | 2 | The comma and exclamation mark separate or surround words. |
don't stop | 2 | The straight apostrophe stays inside the first word. |
don’t stop | 2 | The curly apostrophe follows the same rule. |
well-known author | 2 | A standard hyphen joins the two parts of the compound. |
well–known author | 3 | This uses an en dash (–), which separates words. |
red—blue | 2 | An em dash (—) separates the two words, even without spaces. |
red / blue | 2 | The slash is not a word. |
2026 plan | 2 | A standalone number counts as a token. |
3.14 | 2 | The decimal point splits two sequences of digits. |
[email protected] | 3 | The @ sign and full stop split the address into three parts. |
2026-10-19 | 1 | The standard hyphens keep this date in one token. |
🐈 | 0 | An emoji alone contains no letters or numbers. |
你好世界 | 1 | Adjacent letters form one token; this rule does not identify Chinese word boundaries. |
The decimal, email and Chinese examples show the limits of a simple counter. For language-specific segmentation or a formal submission limit, check the system that will receive your work.
A reusable 15-word test
Copy this complete block into the Word Counter. The expected total is 15 words. Blank lines do not add words. Keep the punctuation as shown: changing a dash to a hyphen can change the result.
Hello, world! don't stop well-known author red—blue 3.14 [email protected] 2026-10-19 🐈 你好世界
Download the plain-text test · Open Word Counter
This sample can be reused in a lesson, a writing guide or a software test. If you explain these results elsewhere, this section provides a stable source link: LetterToolkit word-count test and counting rules.
Why Microsoft Word, Google Docs and websites can disagree
Different products make different choices for compounds, em dashes, URLs, email addresses, emoji, symbols and non-Latin scripts. Some systems are optimized for publishing conventions while others use simpler token rules. A difference of one or two words does not automatically mean one counter is broken; it can mean the products use different definitions.
When a count matters for a formal limit, treat the destination system as the final authority. Draft with a consistent counter, then verify the final document in the submission platform.
Apostrophes and contractions
English contractions such as can't, we're and they've are normally understood as single written words, so keeping the apostrophe inside the token produces an intuitive result. Possessives such as student's are handled the same way.
Curly apostrophes copied from a word processor should behave like straight apostrophes. This is one reason a Unicode-aware counter is preferable to a rule that only understands basic ASCII letters.
Hyphenated words
There is no single convention that fits every context. LetterToolkit treats a common hyphenated sequence such as part-time as one word. A publisher may choose to count it differently. If you are working to a strict academic or editorial standard, check that standard before relying on any online counter.
Numbers, dates and mixed text
Numbers can carry meaning just like alphabetic words, so a standalone number such as 2026 is counted as a token. Punctuation around it normally does not add another word. Complex strings such as URLs, source-code snippets or formulas can be interpreted differently by different systems, so those are good cases for a final destination check.
Sentences and paragraphs are separate measurements
A word total does not tell you how readable a passage is. Sentence count, average sentence length, paragraph count and repeated-word frequency can reveal structure that a raw total cannot. LetterToolkit's Text Analyzer combines those measurements when you need more than a single number.
A repeatable workflow for strict limits
- Write and edit normally rather than cutting useful content too early.
- Use one counter consistently while drafting.
- Clean accidental repeated spaces or pasted formatting if needed.
- Check the final total in the system that enforces the limit.
- If the counts differ, inspect hyphenated terms, unusual punctuation, URLs and numbers first.
Try the rule on your own text
Open the Word Counter to see words, characters, sentences and paragraphs update as you type. The detailed methodology for LetterToolkit's tools is documented on How LetterToolkit Works.
Reviewed and examples checked: October 19, 2026.