Text stats: words, reading time, and limits

Published: 2026-09-05

How word, character, line, and paragraph counts work; why reading-time estimates depend on words-per-minute; and when platform limits make those numbers matter.

Drafts, bios, SEO snippets, and support macros all live under limits—characters for a tweet or meta description, words for a brief, lines for a paste into a ticket. A text statistics pass answers “how long is this?” without opening a full word processor. Online counters often upload the paste; for unpublished copy, prefer a local tool. LocalTools’ Text statistics counts in the browser and never sends the text to a server.

What each count means

Metric	Typical definition (this site)	Watch out for
Characters	Length of the string as the browser sees it (`String.length`, UTF-16 code units)	Emoji and some symbols are more than one code unit; “grapheme” length can differ
Characters (no spaces)	Same string with all whitespace removed	Tabs and newlines count as spaces here
Words	Tokens after trim, split on whitespace	Hyphenated compounds and CJK text do not always match “dictionary words”
Lines	Segments split on `\n` / `\r\n`	Trailing newline still adds a line; empty input is zero
Paragraphs	Non-empty blocks separated by a blank line	Soft wraps in a UI are not paragraph breaks
Sentences (estimate)	Split after `.` `!` `?` `…` plus following space	Abbreviations (`Dr.`, `U.S.`) and lists without terminal punctuation skew the count

These heuristics are English-oriented and good enough for drafts and limits. They are not a linguistic analysis of every script.

Words vs characters vs “platform limits”

Different products enforce different units:

Characters — X/Twitter-style posts, SMS segments (with encoding caveats), many form fields, Open Graph / Twitter Card descriptions, and HTML meta description length guidelines.
Words — editorial briefs (“800-word post”), reading-level estimates, and some LMS or exam instructions.
Lines — commit messages (git subject + body), code review comments, and “keep it under N lines” paste quotas.
Bytes — rare in UI copy tools, but APIs and URL length caps care about encoding; character ≠ byte for non-ASCII.

When a limit is stated in characters, trust the character (and “no spaces”) rows—not the word count. When an editor asks for word count, ignore character-based SEO folklore until you convert intentionally.

Reading time is an estimate, not a clock

Reading time is usually:

minutes ≈ word count ÷ words per minute (WPM)

Silent adult reading of general English prose is often cited around 200–250 WPM. Technical docs, dense legal text, and non-native readers are slower; skimming headlines is faster. The Text statistics tool defaults to 200 WPM and lets you change the rate so the estimate matches your audience—not a universal truth.

Treat the result as an informal “about N minutes” label for blogs and docs. It does not model images, code blocks, or re-reads. Sentence count is likewise a rough heuristic for structure, not a grade-level score.

Practical workflow

Paste the draft (or a section under a hard limit).
Check the unit your platform uses: characters vs words vs lines.
Adjust reading speed if you are labeling an article for readers.
Copy summary into a ticket or brief when you need a plain-text report.
If whitespace is noisy (mixed CRLF, trailing spaces), normalize first with Whitespace normalizer, then re-count.
For filler mockups while you design layout, generate placeholder prose with Lorem ipsum and measure that instead of real copy.

For URL-safe titles and filenames after you settle length, see Slugify and text case styles.

Why count locally

Paste boxes that “analyze reading level” or “optimize SEO” on a remote server see your full draft—including unreleased product names and customer quotes. Local counting keeps that text in the tab. LocalTools computes statistics string by string in your browser; nothing is uploaded for analysis.

Try it locally

Open Text statistics:

Paste or type in Your text—counts update as you edit.
Set Reading speed (words per minute) to match silent reading, technical docs, or a slower audience.
Use Copy summary for a plain-text report of characters, words, lines, paragraphs, sentences, and estimated reading time.