People ask for strange things: a WhatsApp message with nothing in it, an empty line in an Instagram bio, a game name that looks blank. Type a few spaces and the app refuses or removes them.
The reason is simple. Almost every app trims input before it saves it. If nothing is left after the trim, the input counts as empty. The interesting part is what "nothing" means, because trim functions follow Unicode rules, and some characters that look blank are not whitespace by those rules.
What trim() actually removes
In JavaScript, String.prototype.trim() removes characters the spec calls WhiteSpace or LineTerminator. That covers every character in the Unicode "Zs" (space separator) category, plus tab, line breaks and U+FEFF.
Python's str.strip() with no arguments removes characters for which str.isspace() is true. That is close to JavaScript, with one difference worth knowing: Python does not treat U+FEFF as whitespace.
Here is what each runtime does with some characters that look blank:
| Character | Code | Category | Removed by JS trim() | Removed by Python strip() |
|---|---|---|---|---|
| Space | U+0020 | Zs | yes | yes |
| No-Break Space | U+00A0 | Zs | yes | yes |
| Figure Space | U+2007 | Zs | yes | yes |
| Narrow No-Break Space | U+202F | Zs | yes | yes |
| Zero Width No-Break Space (BOM) | U+FEFF | Cf | yes | no |
| Zero Width Space | U+200B | Cf | no | no |
| Word Joiner | U+2060 | Cf | no | no |
| Hangul Filler | U+3164 | Lo | no | no |
| Hangul Choseong Filler | U+115F | Lo | no | no |
| Braille Pattern Blank | U+2800 | So | no | no |
You can check it yourself:
for (const cp of [0x3164, 0x2800, 0x115F, 0x200B, 0xA0, 0x202F, 0xFEFF]) { const c = String.fromCodePoint(cp); console.log(cp.toString(16), /\s/.test(c), c.trim().length === 1); } // 3164 false true <- not whitespace, survives trim // 2800 false true // 115f false true // 200b false true // a0 true false <- whitespace, trimmed away // 202f true false // feff true false import unicodedata for c in '\u3164\u2800\u200B\u00A0\uFEFF': print(f'U+{ord(c):04X}', unicodedata.name(c), unicodedata.category(c), 'survives strip:', len(c.strip()) == 1) Why the survivors look blank
Hangul Filler (U+3164) is a letter (category Lo). In Korean text it holds the place of a missing part of a syllable block and is normally drawn with no ink. Because it is a letter, trim leaves it alone, and the app sees one character of content.
Braille Pattern Blank (U+2800) is a symbol (category So): the braille cell with no dots raised. It renders as an empty space with real width.
Zero Width Space (U+200B) also survives trim, but it has no width at all. Some apps strip format characters (category Cf) on top of trimming, or check for "visible" content, so a message made only of zero width spaces is often still rejected.
What this looks like in practice
I run a small set of browser text tools, and the invisible character page picks the character based on where you plan to paste it:
- WhatsApp message, WhatsApp name, Discord, blank comment: Hangul Filler (U+3164)
- Instagram empty lines: one Braille Pattern Blank (U+2800) per line
In a test on 9 November 2026, six blank messages sent this way went through in WhatsApp. Apps change their input rules without notice, though, so the page has a "Didn't work? Try another" button that switches to the next candidate character.
If you build the input field
If you want to reject input that only looks empty, trim() is not enough. A stricter check:
const looksBlank = /^[\s\u115F\u1160\u3164\uFFA0\u2800\u200B-\u200D\u2060\uFEFF]*$/u; looksBlank.test('\u3164'); // true looksBlank.test('\u2800\n\u2800'); // true looksBlank.test(' \u200B '); // true looksBlank.test('hi\u3164'); // false U+1160 (Hangul Jungseong Filler) and U+FFA0 (Halfwidth Hangul Filler) are in there because they behave like U+3164.
Two cautions:
- Only use this to test whether a whole field is blank. Do not use it to strip characters from real text: U+200C and U+200D are a normal part of emoji sequences and several scripts.
- Fillers are legitimate Korean characters. Blocking them inside text would break real input. Blocking a field made of nothing else is a product decision, not a correctness fix.
Summary
-
trim()andstrip()remove Unicode whitespace, not "everything that looks empty". - Hangul Filler and Braille Pattern Blank are not whitespace, so they survive and render blank.
- Zero width characters survive trim too, but many apps reject them separately.
- If blank-looking input is a problem in your app, test the whole value against an explicit list.
Have you seen an app that blocks Hangul Filler, or one that accepts zero width spaces? I'd like to know which.



