Word Count Explained: Words vs Characters, Reading Time, and Platform Limits
A practical guide to word count: how words differ from characters, how reading time is estimated, how CJK text is counted, and the real character limits for Twitter, SEO meta tags, and more.
Word Count Explained: Words, Characters, Reading Time, and Platform Limits
Word count looks like one number, but the moment you start writing for real platforms it splits into several. A blog editor wants words. Twitter wants characters. Google wants a meta description under a pixel budget. A subtitle line wants characters per second. The same paragraph can be "too long" and "too short" at the same time depending on which ruler you hold up to it.
I built and use the Word Counter for exactly this reason — to stop guessing which limit a draft is about to break. This guide walks through what each metric actually measures, where the common platform limits come from, and the one place most counters quietly get it wrong: Chinese, Japanese, and Korean text.
Words vs Characters: They Are Not the Same Ruler
A word count splits text on whitespace. A character count tallies every keystroke — letters, spaces, punctuation, emoji. They diverge fast.
Take this sentence:
The quick brown fox jumps over the lazy dog.
- Words: 9
- Characters (with spaces): 44
- Characters (without spaces): 36
- Sentences: 1
That gap between 44 and 36 is the eight spaces. It matters more than it looks. Twitter, SMS, and most "max length" form fields count characters with spaces. Some database fields and old systems count without. When a field rejects your text and you swear it should fit, the with/without-spaces distinction is usually the culprit. That is why a good counter shows both side by side instead of forcing you to count by hand.
Reading time is a third ruler entirely, and it is derived from words, not characters — at least for English.
How Reading Time Is Actually Calculated
Reading time is an estimate, and the honest version of it tells you its assumption. The widely-cited average for adult silent reading is 200 to 250 words per minute (a 2019 meta-analysis by Brysbaert in the Journal of Memory and Language put the mean silent reading rate around 238 wpm for English non-fiction).
The Word Counter uses 200 wpm for English. So a 1,000-word article reads in about five minutes. A 600-word newsletter is roughly three. The point of the estimate is not surgical precision — it is to catch the draft that has quietly ballooned to a 12-minute read when you promised a "quick update."
CJK text needs a different rate, which is where most tools fall apart.
How Chinese, Japanese, and Korean Get Counted
English has spaces, so splitting on whitespace works. Chinese does not put spaces between words, so a naive whitespace splitter sees one giant "word" and reports a word count of 1 for an entire essay. That is wrong, and it is a surprisingly common failure.
The convention that actually matches how people think — and how WeChat and Weibo display their limits — is to count each CJK character as one word. So 你好世界 is four "words," not one. The Word Counter detects CJK characters and applies this rule, while still splitting Latin words on whitespace, so a mixed English-and-Chinese draft counts both correctly in one pass.
Reading time also changes. CJK reading is measured in characters per minute, not words per minute, because each character carries more meaning than a single Latin letter. The tool uses 300 characters per minute for CJK silent reading, another commonly-cited average. If you write the kind of bilingual product copy where an English headline sits above a Chinese subhead, this is the difference between a believable reading estimate and a meaningless one.
The Platform Limits Worth Memorizing
Most of the time you are not chasing a target word count — you are trying to fit under a hard ceiling. The numbers below come up constantly:
- Twitter / X post: 280 characters (raised from the original 140 in 2017). Verified accounts can go far higher, but 280 is still the line most replies and quote-posts are measured against.
- SEO meta description: Google does not enforce a character count; it truncates by pixel width, roughly 920 pixels, which works out to about 155–160 characters for typical Latin text. Past that, your description gets cut with an ellipsis in the search result.
- SEO title tag: about 580 pixels, or 50–60 characters, before truncation.
- SMS: 160 characters per single message segment (GSM-7 encoding); longer messages split into concatenated parts.
- Instagram caption: 2,200 characters, but only the first ~125 show before "more."
- Meta / OpenGraph description: keep it under ~200 characters for clean previews.
Notice that the two SEO numbers are pixel limits, not character limits — a title full of wide capital letters truncates sooner than one full of thin lowercase. Character count is your best proxy, but it is a proxy. When I draft a meta description, I aim for the 150–155 character band so there is room before the cut. The Meta Description Brief Generator is what I reach for when I want to draft to that target on purpose instead of trimming after the fact.
Reading the Numbers Without Tunneling on One
The trap with any counter is staring at a single figure. Word count alone tells you nothing about whether the text is dense or skimmable. A 500-word block with no paragraph breaks reads slower than 700 words split into six paragraphs, even though the smaller number "looks" tighter.
So when I review a draft I read four numbers together: word count for the editorial brief, character count for the field limit, paragraph count for density, and reading time for the reader's patience. If the paragraph count is low relative to words, the text needs breaking up regardless of what the word total says. If you are doing the opposite — auditing which words you actually overuse — pair the counter with a Word Frequency Counter to surface the term you have leaned on twelve times in 400 words. And when you are reshaping copy to fit a title or heading style, the Case Converter handles UPPERCASE, Title Case, and slug forms without retyping.
Counting Without Sending Your Draft Anywhere
One more thing that gets ignored: privacy. A word counter does not need a server. The counting is arithmetic over the string you paste — splitting on whitespace, matching CJK ranges, dividing by a reading rate. The Word Counter runs entirely in the browser, so an unpublished post, a client deliverable, or an internal memo never leaves your machine. There is no reason to upload a private draft to a server-side tool just to learn it is 1,240 words.
The next time a form rejects your text or an editor asks "how long is this," you will know which of the four rulers is the one that matters — and you will have all of them in front of you at once.
Made by Toolora · Updated 2026-06-13