Detect zero-width spaces, NBSP, and full-width whitespace
Invisible Unicode characters frequently slip into copied prompts, API payloads, YAML files, and localized text. Two strings look identical on screen while exact equality checks or downstream parsers fail.
Common invisible & lookalike whitespace characters
| Character | Code point | Visual appearance | Where it comes from |
|---|---|---|---|
| Ordinary space | U+0020 |
Normal gap | Standard keyboard input |
| No-break space (NBSP) | U+00A0 |
Identical gap | Rich text editors, HTML, Word |
| Zero-width space (ZWSP) | U+200B |
Zero width (invisible) | Web copy-paste, PDF extraction |
| Ideographic full-width space | U+3000 |
Wide CJK gap ( ) |
Japanese/Chinese IME input |
| Byte order mark (BOM) | U+FEFF |
Invisible prefix | Windows UTF-8 CSV/script exports |
Inspect exact code points locally
- Open the working invisible-character example below.
- Turn on Show whitespace to reveal markers (
·for space,→for tab,□forU+3000,◦for NBSP/ZWSP/BOM). - Click any highlighted marker to read its exact Unicode code point, character name, and UTF-16 column offset in Inspect.
Open the invisible-character inspector
Check UTF-8 hygiene across teams
When comparing Japanese (全角 vs. 半角), Vietnamese, or English payloads, always keep Ignore whitespace disabled first so hidden separators remain flagged. Never blindly strip all non-ASCII whitespace before checking whether a target system actually expects NBSP or ZWJ (U+200D in emoji/scripts).
folio