An invisible character remover deletes a pinned set of Unicode scalars and tells you how many of each it deleted. The rest of the text stays, including ordinary spaces, tabs, newlines, and no-break spaces. A run that finds none of the selected marks is still a success, with a removed count of zero and the original text unchanged.
The default set is on unless you clear it. It contains U+00AD soft hyphen, U+180E Mongolian vowel separator, U+200B zero-width space, U+2060 word joiner, and U+FEFF zero-width no-break space, also used as a byte-order mark. These marks can sit inside a word and are easy to paste in from a web page or a document without seeing them.
A second set is off unless you turn it on: U+200C zero-width non-joiner and U+200D zero-width joiner. Those two scalars can change how Arabic, Persian, Indic, and other scripts shape. Removing them silently would glue or split letters the author meant to keep apart. The checkbox exists so that choice is visible. Leave it off when you are cleaning Latin text and you are not sure the paste uses those scripts.
The Whitespace Cleaner is a different operation. It can collapse ordinary spaces, including no-break space, and it can turn line endings into line feeds. It does not delete U+200B, U+200C, U+200D, U+2060, or U+00AD inside a word. It treats U+FEFF as whitespace and can collapse it into a space. This page deletes U+FEFF instead. Use the cleaner for spaces and blank lines. Use this page for the named invisible marks.
If a mark is still in the text because this page is specified to leave it, the Unicode Code Point Inspector can name the scalar. Bidi controls such as U+200E and U+200F are in that group. This page does not remove them.
Both sets off is an error, because a run with an empty deletion list would pretend to clean and change nothing without saying why. Empty input is also an error. A paste that contains none of the selected marks is not an error.
The example value is the letters h, e, l, a zero-width space, l, o. With the default set on and joiners off, the result is hello and the zero-width space row shows a count of 1. The same letters with a soft hyphen, a Mongolian vowel separator, a word joiner, or U+FEFF in the middle also collapse to the surrounding letters, and each of those scalars is counted on its own row.
A zero-width non-joiner or joiner stays in the text while the joiner option is off. Turn the option on and those two scalars are deleted and counted. A normal space, a tab, a newline, a no-break space, U+200E, and U+200F stay in every mode this page offers.
hel + U+200B + lo → hello
removed U+200B: 1
Deleting these marks can change how a line wraps or how a word searches, and it can change shaping when the joiner option is on. It does not prove the text is safe to render, and it does not prove two strings are the same word in every script. If the question is “what code point is this mark,” inspect it. If the question is “collapse the spaces,” use the whitespace cleaner.
Removal runs in the browser. The paste is not uploaded. Copy writes the cleaned text to the clipboard only when you use Copy, through the same fallback the other text tools use when the clipboard is blocked. Do not paste secrets you would not want left in the field.
| Code point | Name | Removed |
|---|
The joiner option is off by default. U+200C and U+200D can change the shaping of Arabic, Persian, Indic, and other scripts, so they are not removed unless you turn that option on. Ordinary spaces, tabs, newlines, no-break spaces, and bidi controls stay. This is not a security sanitizer.
U+200C and U+200D can change the shaping of Arabic, Persian, Indic, and other scripts. They are removed only when you turn that option on.
No. If none of the selected marks are present, the result is success, the count is 0, and the text is unchanged.
No. U+200E, U+200F, ordinary spaces, tabs, newlines, and no-break spaces stay. The whitespace cleaner is the tool for ordinary space cleanup.
No. It deletes one named list of scalars. It does not detect homoglyphs, normalize the text, or claim the result is safe.