What this tool does

The Word Counter & Text Statistics tool reads any text you paste and breaks it down into nine measurements: characters with spaces, characters without spaces, words, sentences, paragraphs, lines, syllables, estimated reading time, and estimated speaking time. Below the counts you also get a character-frequency chart showing the 30 most common non-whitespace characters as horizontal bars.

It's designed for the moments when "roughly 500 words" isn't good enough, submission limits, social-media caps, SEO meta-description targets, dialogue pacing for scripts, syllable counts for poetry, or reading-time estimates for articles and emails.

How to use it

1. Paste or type your text into the input area. Click Sample if you want to see real numbers fast. 2. The KPI grid updates automatically about 180 ms after you stop typing, no Run button needed. 3. Hit Copy JSON to grab the entire stats object (all nine fields plus the frequency table) for use in another tool, a spreadsheet, or a teammate's review.

The character-frequency bars are sorted by count, descending. Spaces are excluded; everything else, letters, digits, punctuation, emoji, is counted as it appears (a multi-part emoji such as 👍🏽 is one entry).

How the counts are defined

Counting rules differ between tools. Knowing exactly how each metric is computed here matters when you're comparing against an editor, a school assignment, or an API limit:

  • Characters, every user-perceived character (Unicode grapheme cluster) in the text, including spaces, tabs, and newlines. An emoji with a skin tone (👍🏽) or an accented letter typed as two code points counts as one character.
  • Characters (no space), same, minus any whitespace character.
  • Words, Unicode word boundaries (UAX #29). Contractions like "don't", decimals like "3.14", abbreviations like "U.S.A." and hyphenated words like "e-mail" each count as one word. Chinese, Japanese and Thai text, which has no spaces, is split into dictionary words.
  • Sentences, Unicode sentence boundaries. Decimal numbers and dotted abbreviations like "U.S.A." do not end a sentence; a title such as "Dr." followed by a capitalised name still can, there's no perfect heuristic.
  • Paragraphs, blocks separated by one or more blank lines.
  • Lines, newline-separated rows, exactly what your editor's status bar shows.
  • Syllables, counted by a vowel-cluster heuristic per word. Accurate enough for English readability estimates; not a substitute for a phonetic dictionary in production NLP.

Reading time vs speaking time

Reading time is computed at 238 words per minute, which is the average for adult silent reading on screens (the figure most readability research converges on). Speaking time uses 150 wpm, the typical conversational or presentation pace.

Both numbers are rounded up to the nearest minute, so a 50-word note still shows "1m". If you need second-level precision, divide your word count by the wpm figure manually.

Use it for…

  • Blog posts and articles, confirm an "8-minute read" estimate before publishing.
  • Twitter / X / Bluesky posts, get a quick character count without losing focus.
  • SEO meta descriptions, keep titles under 60 characters and descriptions under 160.
  • Speeches, podcasts, video scripts, see the speaking-time estimate before recording.
  • Academic essays, exact word counts matched to assignment requirements.
  • Translation work, character-frequency distributions help spot encoding issues (e.g. a stream of ? glyphs after a bad import).

Common gotchas

  • Word count vs Microsoft Word. Word's counter splits on different boundaries, for hyphenated technical terms or em-dashed phrases, the two will disagree by 1–3%. Don't worry about it unless you're under a hard institutional cap.
  • Sentence count for code or markdown. Code, lists and headings without end punctuation don't follow sentence rules, so the count is only a rough guide there. Strip them out if the count matters.
  • Reading-time on short text. Anything under 238 words rounds up to "1m" because the floor is 1 minute, useful as a UI guarantee, surprising as a math result.

Privacy

Every count, every syllable estimate, every frequency bar is computed in your browser. Your text is not uploaded, logged, or persisted server-side. Recent inputs are stored only in your browser's local storage and you can clear individual items from the Recent card at any time.

Frequently asked

Is my text uploaded anywhere?

No. All counting, syllable estimation, and frequency analysis runs in your browser. Nothing is sent to a server. Recent inputs are stored only in your browser's localStorage, and you can remove individual entries from the Recent card.

Why does the word count differ slightly from Microsoft Word?

Word boundaries are not standardised. This tool follows the Unicode word-boundary rules, so 'don't', '3.14' and 'state-of-the-art' each count as one word, while em-dashed phrases split. Word handles some of these cases differently. The two counts will usually be within 1–3% of each other, which is fine for most submission limits.

How is reading time calculated?

Word count divided by 238 words per minute, rounded up to the nearest whole minute. 238 wpm is the published average for adult silent reading on screens. Speaking time uses 150 wpm, the typical conversational pace.

Why does a 50-word note show '1m' reading time?

The reading-time minimum is 1 minute. The floor exists because UI labels like '0m read' look broken. If you need exact precision, divide your word count by 238 yourself.

How accurate is the syllable count?

It uses a vowel-cluster heuristic — every group of consecutive vowels in a word counts as one syllable, with a silent-e adjustment. Good enough for English readability scores like Flesch-Kincaid. Not a substitute for a phonetic dictionary if you need linguistic precision.

What does the character-frequency chart show?

The 30 most common non-whitespace characters in your text, sorted by count. Useful for spotting encoding glitches (a stream of '?' after a bad import), noticing an over-used letter in a constrained-writing exercise, or sanity-checking translation output.

Does the sentence count handle abbreviations like 'Dr.' correctly?

Mostly. Sentences follow the Unicode sentence-boundary rules, so decimals like '1.5', dotted abbreviations like 'U.S.A.' and 'e.g.' followed by lowercase text do not end a sentence. A title such as 'Dr.' or 'Mrs.' followed by a capitalised name can still be counted as a boundary. There is no perfect heuristic without a full NLP parser; for copy-editing-grade counts, check those cases by hand.

Can I copy all the stats at once?

Yes. Click 'Copy JSON' in the toolbar to grab a structured object containing all nine counts plus the character-frequency table. Paste it into a spreadsheet, a chat, or a teammate's PR review.

Related tools