By Haru · Updated · 2026-09-07
Character Count With Spaces — Which Number Is Meant?
Paste the same paragraph into three counters and you can get three different numbers. Nobody is wrong. There are at least three legitimate ways to count text — with spaces, without spaces, and in bytes — and a form that says 500 characters usually does not say which one it means.
That is a decision you have to make before you submit, and it is not always the same decision. An application form and a message field and a page description all tend to mean different things by the same word.
This page is about choosing. Which basis a form probably means, what to do when it is silent, why byte limits behave strangely with Korean and Japanese text, and why an emoji can count as one character in one place and several in another.
Three counts, one piece of text
With spaces is the largest number and the most common default. Every letter, every space between words, and usually every line break counts as one.
Without spaces removes the gaps. On ordinary prose this is roughly fifteen to twenty percent smaller, which is enough to be the difference between fitting and not fitting.
Bytes is a different measure entirely. It counts the storage the text occupies rather than the symbols you see, and for English the two nearly coincide. For Korean, Japanese and Chinese they do not: one character is commonly three bytes, so a 500-byte limit can mean about 166 characters rather than 500.
A fourth number, word count, is not a smaller version of any of these. It counts runs separated by spaces, which is a fair measure for English and close to meaningless for languages that do not put spaces between words.
When the form does not say, aim at the larger number
Most forms say 500 characters and stop there. If you cannot find out which basis is meant, write to the stricter one, which in practice means counting with spaces.
The reasoning is asymmetric. If the form actually counted without spaces, you have written slightly short and nothing bad happens. If it counted with spaces and you wrote to the smaller figure, your text is cut off, and the cut usually lands mid-sentence at the end — the part you spent the most time on.
There is one way to find out for certain, and it takes ten seconds. If the field shows a live counter as you type, paste your text in and read that number rather than any external tool. The field is the only authority that matters, and any counter that disagrees with it is answering a different question.
If there is no live counter, paste in a test string you have counted yourself and see whether it is accepted or trimmed. A short experiment beats a guess.
Byte limits, and why they are strange in Korean and Japanese
Some fields count bytes rather than characters, which is invisible until it bites. Two conventions are in circulation and they disagree.
Modern systems store text as UTF-8, where Latin letters take one byte, most accented letters take two, and Korean, Japanese and Chinese characters take three. A limit expressed in bytes therefore shrinks dramatically for those languages.
Older forms carry a different convention from the era of two-byte encodings, where any wide character counted as two and everything else as one. Plenty of forms still count that way, especially in application systems and messaging, because that is how the rule was written and nobody revisited it.
The practical upshot: if a limit is stated in bytes and your text is not purely Latin, do not assume. Check both figures — our counter shows the real UTF-8 length and the two-byte convention side by side — and if you cannot tell which the form uses, write to whichever number is larger.
Line breaks, emoji, and the other quiet disagreements
Line breaks are the most common invisible difference. Some counters treat each break as a character, some as two on systems that use a carriage return and a line feed, and some ignore them entirely. On a long piece with many paragraphs this can shift the total by dozens. Unicode has a name for the unit people actually see: the extended grapheme cluster, defined in Unicode Standard Annex #29. A family emoji is one cluster, seven code points and twenty-five bytes at the same time, and every counter picks one of those three to report.
Emoji are worse. A plain emoji is usually one character to a person and one code point to a program, but many are built from several code points joined together — skin tones, flags, families. Counted one way a flag is one character; counted another it is two. Anything that counts by code unit rather than by what a reader sees will disagree with your eyes.
Trailing whitespace is the small one that catches people. A stray space or an empty line at the end still counts in most fields, and it is invisible on screen. If you are one character over a limit and cannot see where, look at the end of the text first.
None of these are bugs. They are different reasonable definitions, and the only way through is to use a counter that shows you which definition it applied.
Choosing a counter
Any counter can produce a number. Three things separate one worth using from the rest.
It shows more than one basis at once. A tool that gives you a single figure has hidden the decision this page is about, and you cannot check it against a limit you are not sure of.
It says what it does with the edge cases. Whether spaces, line breaks and emoji count, and which byte convention it applies, should be visible rather than implied.
And it does not send your text anywhere. Counting is arithmetic; it needs no server. If you are pasting a cover letter, a contract clause or anything else you would not post publicly, a counter that uploads it has taken a copy for no reason at all. Ours counts in the browser and says so on the page.
A live count as you type matters more than it sounds, because writing to a limit is editing, not measuring. Watching the number move while you cut is faster than pasting into a separate tab after every attempt.
What the common limits usually mean
Application and admission forms most often mean characters with spaces, and many of them enforce it with a live counter in the field. Where a form is silent, the safest reading is with spaces, because that is the count the field itself is nearly always doing.
Social posts count characters, but each service has its own rules about links and attachments, and some count a link as a fixed length regardless of how long it is. Read the counter in the composer rather than working it out yourself.
Page titles and descriptions for search results are the odd one out: they are not really limited by characters at all, but by how much fits in a fixed width, so a title of wide characters is truncated much earlier than a title of narrow ones. Counting characters gets you close and the display is the real limit.
Subtitles, product fields and database columns are where byte limits hide. If a field rejects your text at a length that makes no sense in characters, count the bytes and the number will usually explain itself.
Sources Unicode Standard Annex #29 — Text Segmentation
Frequently asked questions
Is with spaces or without spaces the default?
With spaces is the more common default, and it is what almost every live counter inside a form is doing. Without spaces shows up mainly in older writing guidelines and in some editorial contexts. If nothing says otherwise, assume with spaces, because that is both the usual convention and the safer direction to be wrong in.
The form does not say which one it means. What should I do?
Write to the count with spaces, which is the larger number, so you fit either way. If the field has a live counter, trust that above any external tool — it is the thing that will actually accept or reject your text. If it does not, paste a test string of a known length and see whether it survives.
How many characters is an emoji?
It depends on who is counting. To a reader it is one. To a program it may be one, two, or several, because many emoji are assembled from multiple code points — a flag, a skin tone, a family group. Counters that measure what a reader sees will say one; counters that measure storage units will say more. If you are close to a limit, replace the emoji and see how much the number moves.
Why does my word processor give a different number?
Because it is answering a slightly different question. Word processors often exclude headers, footers, footnotes and text boxes from the count, sometimes exclude line breaks, and always apply their own definition of a word. For a form limit, count the exact text you are going to paste, not the document it came from.
What is a byte limit and why did my text fail one?
A byte limit measures storage rather than symbols. In modern encoding a Latin letter is one byte but a Korean, Japanese or Chinese character is usually three, so text in those languages hits a byte limit around three times sooner than the character count suggests. Some older forms instead count wide characters as two. Check both figures if the limit is stated in bytes.
Does a counting tool keep the text I paste in?
It depends entirely on the tool, and most do not tell you. Counting is arithmetic that needs no server, so a well-built counter does it in your browser and nothing leaves the page. Ours works that way and says so where you can see it. If a counter does not say, assume the text was sent, and think twice before pasting anything confidential.