Q: Why does my count differ from MS Word by a few words?
MS Word and Google Docs use slightly different tokenizers. Word treats don't as 1 word, this tool does too. But Word counts web/page as 1 word, while whitespace-split counts it as 1 word here too — and other counters split it as 2. Hyphenated well-known is 1 word everywhere. Discrepancies are usually within 1-2% — fine for essay limits, irrelevant for translation pricing.
Q: Does it count code blocks and Markdown syntax?
Yes — the tool counts every character including backticks, asterisks, and code. For prose-only Markdown counts, paste your rendered output (right-click in your editor → "Copy as plain text") instead of the source. Some publishing platforms (Medium, Substack) only count visible text; budget for ~10-15% shrinkage between Markdown source and rendered counts.
Q: How are emojis counted?
As one grapheme each when possible. The skin-tone modifiers and ZWJ (zero-width joiner) sequences in emoji like 👨👩👧👦 (4 people joined by ZWJ) display as one glyph but expand to 7+ Unicode code points — Twitter, Bluesky, and SMS gateways count them differently. SMS specifically charges one segment per ~70 Unicode characters once any emoji appears, so a 70-char message with one emoji costs the same as two English SMS.
Q: What's a normal article length for SEO?
Depends on intent. Backlinko's 2024 analysis put top-ranking pages at 1,447 average words, but that's a correlation, not causation — long-form just tends to cover more sub-topics that attract long-tail searches. Local business pages, product pages, and definitions rank fine at 300-500 words if the intent is transactional. Don't pad copy to hit a word count; Google detects thin filler.
Q: How does it estimate reading time?
ceil(wordCount / 225) minutes. 225 WPM is the Brysbaert meta-analysis median for adult silent reading of plain English. Technical or unfamiliar material runs slower (180-200 WPM); easy narrative runs faster (250-280 WPM). For audio scripts, the tool also shows speaking time at 150 WPM, suitable for conversational pace.
Q: Why is keyword density still a metric?
Google's ranking algorithm hasn't cared about literal keyword density since 2013, but the metric remains useful as a readability check — if one word is >3% of your text, you're probably repeating yourself. The top-10 word panel surfaces this without prescribing it. Write for humans, not for density targets.