Frequently asked questions

No. Every statistic is computed by JavaScript running in your own browser tab. There is no analysis server, and we never see, log, or store what you type. You can verify this yourself: load the page, turn off your internet connection, and everything keeps working. The single exception is the optional AI tab, which is off by default, clearly labelled, and only sends text to a provider you configure with your own API key.

Yes, free — no account, no sign-up, no trial. There is no hard word limit; the practical ceiling is your device's memory. Documents in the hundreds of thousands of words analyse in well under a second on a modern computer, because the work happens locally instead of waiting on a server.

Because there is no universal definition of a "word." Tools differ on hyphenated compounds, numbers, dates, em-dashes, and slashes. WordLab uses whitespace splitting for the headline count — the same convention as major word processors — so differences are usually a fraction of a percent. If a submission has a strict limit, the counter that matters is the one your recipient uses; leave yourself a small margin.

Accurate enough to guide editing, not precise enough to quote to decimals. Four of the six formulas depend on syllable counts, which software can only estimate for English. Different websites use different estimators, so scores commonly vary by a grade level or two between tools. Watch the trend across your edits rather than the absolute number.

Your text auto-saves to your browser's local storage on your own device — nothing is sent to us. It stays until you clear it: press Clear in the toolbar, or wipe the site's data in your browser settings. Note this also means your draft won't follow you to a different browser or computer.

Yes. After your first visit over HTTPS, the app caches itself and works with no connection at all. The only features that need the network are .docx/.pdf import (a parsing library is fetched from a CDN — your file itself still stays local) and the optional AI tab.

Plain text (.txt), Markdown (.md), HTML, RTF, CSV, Word documents (.docx), and PDFs. Drag and drop works too. Files are read directly by your browser; they are not uploaded to a server. Scanned PDFs that contain only images have no extractable text.

Counting works for any language — the tokenizer understands Unicode, so accented letters, Cyrillic, Greek, Arabic, and other alphabets all count correctly. Two caveats: languages written without spaces (Chinese, Japanese) can't be word-counted by any whitespace method, so use the character count; and the readability formulas were designed and validated for English, so treat their scores on other languages as indicative only.

Word count divided by 220 words per minute — a middle value from published research on adult silent-reading speeds. Speaking time uses 130 wpm, a typical presentation pace. Both are planning estimates; dense material reads slower and light material faster.

It's how often a word appears as a percentage of all words. There is no magic target — modern search engines penalise keyword stuffing rather than rewarding a percentage. Use the table to confirm your topic words actually appear and that no accidental filler word dominates. The stop-word filter hides common words like "the" and "and" so the table shows your real vocabulary.

AI models like GPT and Claude measure text in tokens (subword chunks of roughly four characters of English). If you're budgeting prompts or estimating API costs, the token estimate tells you approximately what a model will see. Exact counts vary by model — treat it as within roughly 10–15% for English prose.

Yes — that's what it's for. Just remember that whoever sets a word limit also chooses the tool that enforces it, and tools differ slightly, so leave a small buffer near hard limits. WordLab gives you statistics about your text; the text itself, and responsibility for it, remain entirely yours.

The site may display advertising (such as Google AdSense) to cover its costs. Ads never receive your text — analysis is local, and there is nothing to send. Ad networks may set their own cookies for frequency capping and, where permitted, personalisation; our Cookie Policy and Privacy Policy explain this and how to opt out. WordLab itself uses only local storage on your device for your draft and settings.

Please tell us! Send the details through the Contact page, ideally with a small sample text that reproduces the issue (never send anything confidential), the count you expected, and the browser you're using. Counting-logic reports are the most valuable feedback we get.