Text cleanup · nothing is uploaded

Remove what
you cannot see.

Paste your text to list every invisible code point in it — by name and by number — then take a clean copy.

0
0
Ready to scan

Paste some text to begin.

0characters changed
0character types
0%of text changed

You get the code points, not just a cleaned blob

Most cleaners hand back stripped text and leave you guessing what was in there. This one names every character it found, gives you its Unicode number, and says whether it was removed or replaced. If you are debugging a broken import or a failing string comparison, that list is the answer you actually needed.

Safe mode leaves working characters alone

Some invisible characters do real work: joiners hold emoji sequences together and shape scripts with joined letters. Safe mode keeps those and removes only what has no business in ordinary prose. Strict mode removes them too, which is right for machine-readable identifiers and wrong for text containing emoji, Arabic, Persian or several Asian scripts.

Frequently asked questions

What is a zero-width space?

U+200B, a character with no width and no visible mark. It occupies a position in the string, so a search, a comparison or a database key can fail on text that looks identical on screen.

Where do these characters come from?

Copying between editors and messengers, typesetting tools inserting soft hyphens and word joiners, right-to-left text carrying direction marks, and deliberate marking with Unicode tag characters (U+E0000–E007F).

Does finding them prove the text came from an AI?

No. They appear in text written by people and are absent from plenty of model output. The tool reports what is in your file, not who wrote it.