Detect invisible Unicode characters in text
Find zero-width spaces, bidirectional controls and special characters before sharing a message. Inspect locally without uploading your text.
Limit: 60,000 UTF-16 code units. Your text is not analyzed automatically; press the button.
Inspection summary
Detected groups:
Matches and positions
Text with visible code markers (first 3,000 characters)
Prepare a copy without changing the original
Select the groups to remove from the copy. Nothing is removed by default: some characters are legitimate in Arabic, Hindi, emojis and other languages.
Located characters
See the Unicode code point, category and context without executing HTML.
Care with text direction
Bidi controls can affect display order, but may also be legitimate in multilingual text.
Human review
This does not find every look-alike character or determine whether text is malicious.
Why invisible characters matter
A word, URL or filename can include characters that are difficult to see. Some support typography and writing systems. Others can cause mistakes when comparing strings or reading an address. Finding a control character is not proof of fraud: consider the context.
Educational guide: examples, risks and limits of invisible characters
Frequently asked questions
Does removing characters make text safer?
Not necessarily. Removal may break emoji sequences, writing systems and data that depend on them. Always check the copy before sharing.
Is my pasted text sent to the internet?
No: this inspector processes the text locally. Other site functions, such as visit statistics, can make network connections.