ToolwebWhy text counts differ: Unicode, whitespace and UTF-8 bytes
Two editors can report different counts for the same-looking sentence. The reason is often a different definition of one character, rather than an arithmetic error.
1. Identify the unit in the limit
A 500-character limit, a 500-byte limit and 500 characters excluding whitespace are different requirements. Even “character” can mean code units, code points or visible grapheme clusters. This site’s character total counts Unicode code points.
2. Compare small examples
A uses one code point and one UTF-8 byte. The Hangul character 가 also uses one code point but three UTF-8 bytes; 😀 uses one code point and four bytes. There is no single bytes-per-character multiplier for mixed text.
| Text | Code points | UTF-8 bytes |
|---|---|---|
| A | 1 | 1 |
| 가 | 1 | 3 |
| 😀 | 1 | 4 |
| 가 A | 3 | 5 |
| e + combining accent U+0301 | 2 | 3 |
3. Look for combined representations
An accented é can be represented by U+00E9 or by e followed by combining mark U+0301. They look similar but have different code-point counts. A family emoji likewise joins several code points into one visual symbol.
Normalization changes a representation and is a separate operation. The counter does not silently normalize your text. If you compare text between systems, check their normalization rules while preserving the original.
4. Invisible whitespace still contributes
The non-whitespace total excludes spaces, tabs and line breaks. Lines count real newline separators, not text wrapping caused by a narrow window. A final newline adds an empty line, so inspect the end of copied text too.
CRLF and LF can occupy different byte counts in files. The text comparison tool normalizes those line endings before comparing them. Browser text fields can also normalize newlines, so a pasted text count is not necessarily the exact byte size of the source file.
5. Treat word counts as a stated convention
Word boundaries depend on language and segmentation rules. Space-separated English prose behaves differently from many East Asian writing systems. This tool uses language-aware browser segmentation where available, which is not a guarantee of a publisher’s or translator’s billing unit.
6. Finish in the receiving system
Use one consistent convention while drafting, then verify in the actual submission form. A draft exactly at the limit can exceed it when the destination counts emoji or invisible characters differently. Compare matching paragraphs to separate content edits from whitespace changes.
- Confirm whether the limit is in characters, words or bytes.
- Inspect leading and trailing spaces and empty lines.
- Use the destination’s final count for emoji or combining characters.