Platform limits fall into two kinds, and the difference decides how much you should worry about them. A hard limit refuses the text: a post on X stops at 280 characters and the button will not submit. A soft limit truncates the display: a title tag has no maximum at all, and search results simply cut it off at whatever fits, usually somewhere around 60 characters.
The soft ones cause more damage, because nothing tells you it happened. You write the sentence, it saves, and the second half of it is never shown to anyone.
Which limits are worth writing against?
Four cover most work, and only two of them are enforced.
| Where | Limit | Kind |
|---|---|---|
| Title tag | ~60 characters | Soft — truncated in results |
| Meta description | ~155 characters | Soft — truncated in results |
| Post on X | 280 characters | Hard — refuses to submit |
| Instagram caption | 2,200 characters | Hard, with a soft fold after the first line or two |
The Instagram case has both kinds at once. The 2,200 limit is enforced and almost nobody hits it; the practical limit is the point at which the caption collapses behind a "more" link, which is far shorter and is where the actual writing decision lives.
Why is a title tag measured in pixels?
Because search results are rendered at a fixed width and the letters are not the same width. An i is narrow and a W is wide, so a 60-character title in a condensed sentence can fit where a 55-character title full of capitals does not.
That is why every guideline gives a range rather than a number. Sixty characters is a good working target and it is a proxy for the real constraint, which is roughly 580 pixels of rendered width. Titles built mostly from short words survive longer than the character count suggests; titles with long brand names and capitals get cut earlier.
The practical consequence: put the important words first. Whatever is at the front is guaranteed to be shown, and everything after the first forty characters is a bet.
What counts as a word?
A run of letters or digits separated by whitespace, which means hyphenated and apostrophised words count once. "Well-known" is one word, "don’t" is one word, and "state of the art" is four.
That definition matters when a limit is expressed in words rather than characters. An academic abstract capped at 250 words and a marketing brief capped at 250 words are counting the same way, and a document with many hyphenated compounds will run longer in characters than one without.
Sentence and paragraph counts are more approximate. A full stop ends a sentence, which makes "e.g." and "Dr." into sentence boundaries in almost every counter, including this one — worth knowing before quoting a sentence count in anything that matters.
How reliable is a reading time?
Less than the badge implies. The figure here uses 238 words per minute, the silent reading rate for non-fiction from Marc Brysbaert’s 2019 review of nearly two hundred studies. That review also found a range across studies from roughly 175 to 300 words a minute depending on the reader and the material.
So a five-minute read is somewhere between four and seven minutes for a real person. Treat any reading-time estimate — this one included — as an order of magnitude rather than a promise, and as a signal about length rather than about time.
Speaking time is a different rate again, around 130 words a minute for clear presentation. A thousand-word script is about four minutes to read silently and nearly eight to say out loud, which is the gap that catches people out when a blog post becomes a talk.
What is keyword density actually for?
Diagnosis rather than targeting. Seeing that one word makes up 8 per cent of a piece tells you something about the writing — usually that a phrase has been repeated more than the sentences needed.
It is not a target to hit. Writing toward a density figure produces text that reads as though it were written toward a density figure, which is the specific failure mode search engines have been detecting for two decades. The useful reading is the opposite: if a term you expected to be central does not appear in the list at all, the piece may not be about what you think it is.
Common words are excluded from the list for the same reason. Knowing that "the" is your most frequent word is not information.
Questions people ask
Where do the length limits come from? Platform documentation for the hard ones, and observed truncation for the soft ones. The soft figures drift, which is why they are quoted as approximations and are worth rechecking rather than memorising.
Does the counter include spaces? Both figures are shown, because different platforms count differently. Character limits on social platforms generally include spaces; some legacy systems do not.
How are emoji counted? As single characters here, which matches how most platforms display them and not how all of them count them. A platform counting bytes or code units may charge more for one emoji than you expect.
Is my text sent anywhere? No. Everything is counted in the browser, and nothing you paste leaves the page.
Write to the hard limits and design for the soft ones, because only one of the two will tell you when you have gone over. The word counter shows the counts against each target as you type, the UTM link builder handles the other half of a post that has to fit, and the word unscrambler is next door for when the sentence needs a shorter word.