Free Online Website Word Counter
Count words, characters, and reading time from a webpage.
| Level | Count |
|---|
- Scripts, styles, SVG, and template elements are excluded from the count.
- Navigation, footers, and boilerplate that are part of the body text are included.
- Reading time is an estimate, not a promise of how long a visitor will stay.
The page is fetched by this site's server and stripped of scripts, styles, and markup before counting. The result reflects the HTML the server received, not text injected later by JavaScript.
About the Website Word Counter
Word count is a rough but useful content metric. Editors use it to plan an article, reviewers use it to check a brief, and SEO teams use it to compare a page against competing results. Counting by hand is impractical, and counting in a word processor includes navigation, captions, and cookie banners that a visitor skims rather than reads. This tool fetches a page, strips scripts and styles, normalizes whitespace, and counts the remaining words and characters.
What is measured
- Words: sequences of letters, numbers, apostrophes, and hyphens in the visible body text.
- Characters: the length of the normalized text, including spaces but excluding markup.
- Reading time: words divided by two hundred, rounded up. Average adult reading speed varies widely, so treat it as a label rather than a precise duration.
- Headings: counts of H1, H2, and H3 elements, which hint at structure and scannability.
- Images: total images and how many lack an alt attribute.
The measurement is deliberately honest about its limits. The server returns aggregate counts, not the extracted text, so this page cannot produce a top-words list, keyword density, or readability score without inventing data. Those analyses require the actual body text, and displaying them from counts alone would be misleading. If you need keyword frequency, copy the article into a text analysis tool or use a crawler that exposes page content. What you see here is exactly what was measured and nothing more.
Two pages with the same word count can feel very different. A long page broken into short paragraphs, lists, and descriptive headings is easier to scan than an undifferentiated wall of text, even though the totals match. Reading time also ignores the complexity of the subject and the number of images or code samples. Use the numbers as a starting point, then read the page yourself or ask a few representative users. For editorial planning, compare pages within the same format rather than across a whole site.
The fetch is server-side, which avoids browser cross-origin restrictions but introduces other caveats. Pages that require a login, a cookie consent, or JavaScript rendering may return a shell with very little text. Bot protection can replace the article with a challenge page. In those cases the count will be lower than what a visitor sees. The HTTP status and redirect count help you judge whether the result is trustworthy. Private and local addresses are refused by the endpoint, and only public HTTP or HTTPS URLs are requested.
How to use this tool
- Enter the public URL of the page you want to measure.
- Click Count words and wait for the server to fetch and parse the document.
- Review the word, character, reading-time, heading, and image metrics.
- Copy a plain-text summary or download the report as JSON.
Frequently asked questions
Why is there no keyword density list?
The server endpoint returns only counts, not the page text. A frequency table built from counts would be fabricated. Use a tool that has access to the extracted article text if you need keyword analysis.
Does the count include menus and footers?
Yes, if they are part of the body text. The parser removes scripts, styles, and similar non-visible elements but does not distinguish an article from site chrome. The number is therefore closer to total visible text than to article-only length.
Why does a JavaScript-heavy page show so few words?
The fetcher reads the initial HTML response and does not execute scripts. Content rendered on the client after load is missing. Use a browser-based reading view or a rendering crawler for those pages.