This Korean word counter takes each 어절, the run between two spaces, as one word, particle and all. It gives 글자 수 with and without spaces and Hangul characters alone.
Read the full guide
A cat, a keyboard, 20 words
The sample counts 20 words, 82 characters, 63 without spaces, 58 of them Hangul. The four loose consonants the cat typed, ㅁㄴㅇㄹ, count as four Hangul characters and, stuck to 입니다, as part of one word. Paste your own text over it. Limit for words (어절) checks a limit set in words; a limit in 글자 수 is read off the character lines.
Word counter Korean unit: the 어절
A word here is the 어절, particle included: 학교에 간다 is two words. The count trusts your spacing: 할 수 있다 is three words and 할수있다 is one.
Tips
세 줄 is two words and 3줄 is one: spelling rules let a unit noun close up after a digit.
For a rough count of manuscript paper (원고지) sheets, divide Characters (글자 수) by 200.
If Code points (코드 포인트) far exceeds characters, Hangul is stored as separate letters: two or three times the bytes.
Questions and answers
Are Korean particles counted as separate words?
No. The counter splits only at spaces, and a particle is written onto the word before it, so 한나가 책을 읽는다 is three words. Korean grammar also recognises 가 and 을 as words, so a teacher who counts that way gets five.
Should I count Korean in words or in characters?
Whichever the limit names. For 글자 수, read Characters (글자 수) when it says 공백 포함, spaces included, and Without spaces (공백 제외) when it says 공백 제외. A limit in words means Words (어절). The word counters hub lists the other languages.
Does ㅋㅋㅋ count as a word?
As one word when it stands between spaces, and as three Hangul characters and nine bytes either way. Stuck to a word, as in 진짜ㅋㅋ, it adds characters and no word. The carets of ^^ are symbols: two characters, no word, no punctuation.
글Characters (글자 수)The whole text, character by character, spaces and line breaks too: the figure a Korean limit means when it says 공백 포함, spaces included. A syllable block is one character however many letters are stacked in it, so 한, built from three, counts once and 비빔밥 is three. A block stored as separate conjoining letters still counts once. Manuscript paper (원고지) measures nearly the same way: one Hangul character to a square, digits and Latin letters two to a square, and 200 squares to the usual sheet, which puts 1,000 characters at about five sheets.
가나Without spaces (공백 제외)The figure for a limit marked 공백 제외, spaces excluded: spaces, tabs and line breaks all drop out. Spacing is reckoned one of the hardest parts of Korean spelling, and this is the count that does not mind how you did it. 할 수 있다 and 할수있다 are both four. The sample reads 63 here against 82 with its spaces.
에Words (어절)Whatever stands between two spaces, the unit Korean grammar calls an eojeol (어절). 학교에 간다 is two and 우리는 오늘 동해로 간다 is four. A particle such as 에, 는 or 을 is written onto the word before it, so it never adds to this number, although Korean grammar recognises a particle as a word: 한나가 책을 읽는다 is three here and five by that reckoning. The count follows your spacing and cannot check it. 할 수 있다 is three and 할수있다 is one, 못 하다 is two and 못하다, which means something else, is one. 한국 대학교 사범 대학 is four, 한국대학교 사범대학 is two, and the spelling rules allow both.
Show all 11 counts (8 more)
Only in Hangul 2
ㅎㅏㄴCode points (코드 포인트)How many Unicode characters the text is stored as. Hangul can be stored two ways. Normally a block is one code point, 한 being U+D55C, and this line matches the character count. Decomposed, the same block is three conjoining letters in a row, U+1112 Hangul choseong hieuh, U+1161 Hangul jungseong a and U+11AB Hangul jongseong nieun, which the screen draws as one. It looks identical and still counts as one character, but it is three code points, or two for a block with no final consonant such as 비. When this line runs far above the characters, the text is decomposed: the sample would read 155 here and not 82. The loose letters of ㅋㅋㅋ are a third kind, compatibility letters that never join up (ㅋ is U+314B Hangul letter khieukh).
한Hangul characters (한글)Korean characters and nothing else. Each syllable block counts once, and so does each letter standing by itself, which makes ㅋㅋㅋ three. Latin letters, digits, hanja, spaces and punctuation are left out, so this is the line that says how much of a mixed text is Korean: PDF로 보냈어요 has five. Finished blocks run from 가 (U+AC00) to 힣 (U+D7A3) in Unicode, 11,172 of them. U+3164 Hangul filler, a blank that Unicode classes as a letter, is counted here as well, and as a word when it stands alone. The sample has 58: 54 blocks and the four loose consonants.
Structure 3
↵Lines (줄)Lines as they were typed, the blank ones as well: that title and its two paragraphs come to five. The sample describes a report of three lines (세 줄) and is itself one, since nobody pressed Enter in it. Pressing Enter after the last character makes a new, empty line, and it is counted.
¶Paragraphs (문단)Each stretch of text between line breaks, with blank lines ignored. On manuscript paper a paragraph is shown by leaving its first square empty. Here only the line break shows it, and a space typed at the start of a paragraph changes the space count and nothing else. A title, a blank line and two paragraphs with a blank line between them make three.
다.Sentences (문장)A sentence ends at a full stop, ? or !, and at a line break. Three Korean habits add a sentence that is not there. A quotation that carries its own full stop or question mark before 라고 or 하고 counts as two, as in 그는 “네.”라고 말했다. The official date form with full stops, 2001. 11. 29., ends a sentence when text follows it, where 2001년 11월 29일 does not. And an ellipsis typed as full stops ends one, where the ellipsis character (U+2026 horizontal ellipsis) does not. A row of ㅋㅋㅋ ends nothing.
What the text is made of 2
·Punctuation (문장 부호)Full stops, commas, question marks, quotation marks, brackets, and the dot Korean sets between listed items: 금·은·동메달, typed with U+00B7 middle dot, has two. The corner brackets that go round the title of a book count one each, so 『제목』 has two (U+300E and U+300F). The tilde Korean uses for a range, as in 9월 15일~9월 25일, is a symbol to Unicode (U+007E tilde) and is not counted, and neither are the carets of ^^ (U+005E circumflex accent).
␣Spaces (공백)Spaces, tabs and line breaks, one apiece. On a single line of Korean with one space between words this is the word count less one, 19 against 20 in the sample. Hangul was written without spaces for most of its history. They first appeared in a primer for foreigners in 1877 and spread with the newspaper 독립신문, founded in 1896. The difference between this line and nothing can be the meaning: 아버지가 방에 들어가신다 has father going into the room, and 아버지 가방에 들어가신다, with the same two spaces, has him going into the bag.
How the text is stored 1
가Bytes in UTF-8 (바이트)Every Hangul character is three bytes in UTF-8 and a space, a digit or an ASCII mark is one, so the sample is 58 threes and 24 ones, 198 bytes. EUC-KR, the older Korean encoding, stores a Hangul character in two bytes and ASCII in one, which would make the sample 140. A byte limit that charges two for a Korean character is using that older figure, and this line will run ahead of it. Decomposed Hangul costs six bytes a block, or nine when the block has a final consonant, and the decomposed sample is 417.