ChatGPT zero-width space remover

The tool

What the ChatGPT zero-width space remover does

Paste text copied out of ChatGPT and this takes out every zero-width space and the other invisible characters it knows. U+200B is the one people search for, but it rarely travels alone, so the whole set is handled in one pass rather than that single codepoint.

It runs in your browser. Nothing is uploaded, there is no account, no run limit and no length cap, because no server is involved in the work.

The set

Which invisible characters get removed

Fourteen, and here they all are. Eleven are deleted. Three are space look-alikes and become an ordinary space, because the gap between the words is real even when the codepoint is wrong.

U+200B zero-width space

Occupies a position and renders nothing. The character this page is named for.

U+200C, U+200D, U+2060

Zero-width non-joiner, zero-width joiner and word joiner. Joining and line-break controls.

U+FEFF byte-order mark

Legitimate at the head of a file. Stranded mid-paragraph it breaks parsers and feeds.

U+00AD soft hyphen

Invisible until a line wraps there, then a hyphen appears in a word that never had one.

U+00A0, U+202F, U+200A

No-break space, narrow no-break space and hair space. Swapped for a plain space.

U+202A to U+202E

Five bidirectional controls: two embeddings, a pop and two overrides.

How it works

How to remove zero-width spaces in three steps

Nothing to preview first, because the problem is invisible until the scan runs.

Step 1

Paste the text straight from the chat window. Retyping it first removes the characters, so the scan comes back empty on text that was not clean.

Step 2

Run the removal. It happens instantly, on your own machine.

Step 3

Copy the result. Every visible character comes back exactly as it went in.

Alternatives

Other ways to strip a zero-width space

This is not the only route, and knowing the others tells you when to trust a result.

Plain-text paste

Paste into Notepad or TextEdit in plain-text mode and copy back out. This catches a lot of what arrives from a browser copy, though it also flattens your formatting.

Editor find-and-replace

Word, Google Docs and most code editors accept a Unicode escape in a search box. You need one pass per codepoint, and you need to know which codepoints to look for.

A script

A few lines of Python or JavaScript with a character class. Accurate, repeatable, and overkill for one paragraph.

This page

One pass over all fourteen, with a count of what changed, and no formatting loss.

Scope, stated plainly

What this zero-width space remover will not do

Unicode holds more invisible characters than any one cleaner handles, and a tool claiming to remove all of them is usually wrong. U+2009, the thin space, is named by other pages and is not in this set. Neither are variation selectors U+FE00 to U+FE0F, the Mongolian vowel separator U+180E, the invisible math operators U+2061 to U+2064, or the directional isolates U+2066 to U+2069.

This route also leaves visible characters alone, which is the boundary competitors blur. Em dashes, en dashes, smart quotes, ellipses, asterisks and Markdown headings are all untouched here. It does not collapse runs of spaces or blank lines either. Separate routes on this site handle each of those, and keeping them apart is what stops a cleanup from quietly becoming an edit.

Consequences

What a zero-width space actually breaks

Inside a heading, a slug or a product code, U+200B splits the string for anything comparing text character by character. Two values that look identical stop matching, so a lookup returns nothing and a find-and-replace misses a phrase you can see on screen. In a spreadsheet import the row simply fails to join.

In code, one inside an identifier or a string literal is a syntax error with no visible cause. In a CMS, a stray byte-order mark shows as a glyph in a feed or trips a validator.

There is a security side too. The two directional overrides in the list above change how text displays without changing what it is, so a link or a file name can read differently from what it does. Invisible characters are also a known way to hide instructions in text that a model reads and a person does not.

Checking the result

How to tell it worked

The run reports how many characters were removed or changed, which is the direct answer. If you want an independent check, compare the character count of the text before and after in any counter: it should fall by exactly the number of characters that were deleted. Word count should not move unless a zero-width space was standing in for a real space.

Running it twice is the other useful check. A second pass on already-clean text reports nothing, because there is nothing left in the fourteen to find.