A word counter answers a deceptively simple question — how long is this text? — across the units that matter in different contexts: words, characters (with and without spaces), lines, and paragraphs. Writers check article length, students verify essay limits, marketers fit copy into ad slots, and developers validate content against database column sizes and API limits. All counting happens locally in your browser as you type.

The interesting part is that 'how many characters' has more than one correct answer, and platforms disagree: what Twitter/X counts, what your database column stores, and what String.length returns in JavaScript can be three different numbers for the same visible text.

Words are counted by splitting on whitespace runs — the same definition wc -w uses. This differs from Microsoft Word's counter by small amounts on edge cases (hyphenated compounds, em-dashes, URLs): if a submission limit says '500 words in Word', expect ±1–2% between counters. Lines are newline-delimited; paragraphs are blocks separated by blank lines.

Characters are subtler. JavaScript's String.length counts UTF-16 code units: '🎉'.length is 2, and a family emoji with skin tones can be 11. Unicode-aware counting (code points, or grapheme clusters — what users perceive as one character) gives smaller, more human numbers. Databases add a third dimension: a VARCHAR(255) in MySQL utf8mb4 counts code points, but an index limit or a legacy latin1 column counts bytes — Thai text at 3 bytes/char hits byte limits three times sooner than ASCII.

• X/Twitter: 280 — but weighted: CJK counts double, URLs are flat 23 regardless of length.
• Google search snippets: titles truncate around 50–60 characters, descriptions ~155–160 — pixel-width based, so these are guidelines, not hard counts.
• SMS: 160 GSM-7 characters — but one emoji or Thai character switches the whole message to UCS-2, dropping the limit to 70 per segment.
• Meta descriptions, app store listings, LinkedIn posts, README badges — every platform picks its own unit; when it matters, test in the target.
• Essay/word limits: usually whitespace-token counts; hyphenation and citations are where counters disagree.

Reading-time estimates divide word count by an assumed pace — commonly 200–240 words/minute for adult silent reading of English prose (slower for technical content, ~150–180 for reading aloud, which matters for scripts and speeches). A 1,200-word article is a '5–6 minute read'; a 10-minute conference talk is roughly 1,300–1,500 spoken words. Treat these as planning numbers, not physics.