How We Count: Methods and Standards
This page explains how the tools on Geuljasu count characters and bytes, using only rules confirmed in the site's own calculation code. Service limits are listed only when confirmed in official documents, with the source and the check date (2026-10-30); values we could not confirm are left out.
How characters are counted
- One character is what you see on screen (a grapheme, via Intl.Segmenter). Emoji, combined characters and Hangul stored as separate jamo count as one character each. In browsers without this feature, code points are counted instead.
- The count with spaces counts spaces, tabs, full-width spaces and line breaks as one character each. A Windows line break (CRLF, \r\n) is also one character. The count without spaces drops every character that is whitespace (including tabs, line breaks and full-width spaces).
- The word count is the number of chunks left after trimming the text and splitting on whitespace. The line count is the number of pieces split by line breaks (CRLF, CR or LF), and an empty text has 0 lines. The paragraph count is the number of lines that have content.
- Sentences are counted at . ! ? followed by a space or the end of the text, and after 。!?. A dot in 3.5 or v1.2 is not a sentence end, and fragments with no letter or digit are not counted.
- Manuscript pages (a Korean/Japanese format) are the count without line breaks divided by 200 cells, rounded up. Reading and speaking times use 238 and 150 words per minute on the word count, rounded up to seconds; these are the site's own rough standards.
How bytes are counted
Two values are added up separately for each character. UTF-8 is the length produced by the browser's TextEncoder. EUC-KR is an approximation that counts ASCII (code 0-127) as 1 byte and every other character as 2 bytes.
| Character | EUC-KR approximation | UTF-8 |
|---|---|---|
| Letters, digits, spaces, one line break | 1 | 1 |
| Hangul (가-힣) | 2 | 3 |
| Chinese characters, Japanese kana | 2 | 3 |
| Emoji | Cannot be expressed; this tool counts 2 | 4 |
The text-message type is decided from the EUC-KR approximation: up to 90 bytes is a short message (SMS), up to 2,000 bytes a long message (LMS), and anything longer is MMS. These are the tool's defaults and carriers and sending services can differ. To try it yourself, use the Byte Counter.
The eight limit units
The "limit" feature of the Word Counter and the Character Limit Trimmer measure a limit in one of these units.
| Unit | How it is counted |
|---|---|
| Characters with spaces | 1 per visible character |
| Characters without spaces | 0 for spaces and line breaks, 1 for everything else |
| Words | 1 at the first character of each whitespace-separated chunk |
| EUC-KR approximate bytes | 1 for ASCII, 2 for the rest |
| UTF-8 bytes | The real UTF-8 encoded length |
| X weighted | Code points 0-4351, 8192-8205, 8208-8223 and 8242-8247 count 1; other characters and multi-code-point characters such as emoji count 2 (the 23-character link rule is not applied) |
| Half-width as 0.5 (Weibo style) | Half-width 0.5, full-width 1, line breaks 0 |
| CJK character + English word | 1 per Chinese, Japanese or Korean character, and 1 per chunk of letters or digits (the Word Chinese-edition 字數 method) |
Platform limits: official sources and check dates
Only values confirmed by opening the official documents are listed. Documents can change, so treat the notice on the service's own screen as the final standard.
| Service | What we confirmed | Source | Checked |
|---|---|---|---|
| Google Search results (meta description) | No limit on length; in results it is truncated as needed, typically to fit the device width | Google Search Central "Snippets" | 2026-10-30 |
| X (formerly Twitter) | Up to 280 (weighted). Latin characters and symbols count 1, Chinese, Japanese and Korean characters 2, and any URL counts 23 regardless of length | X developer docs "Counting characters" | 2026-10-30 |
Values such as essay length, text messages, Instagram and YouTube titles were not re-checked against official documents this time, so they are not in this table. The matching presets inside the tools are reference defaults. Related articles: How Many Characters Should a Meta Description Be?, X character limit and Korean, The 90-byte SMS limit and Hangul, Character count with spaces.
Limits to keep in mind
- Each place you submit to may count differently (spaces or not, a line break as one or two characters, bytes). Use the display on the submission screen for the final check.
- All text you enter is calculated only inside your browser and is never sent to a server. The details each tool uses are written under "How this tool calculates" on its own page.
- If a calculation rule changes, this page is updated together with the check date.