Unicode Confusables Detector
The Unicode Confusables Detector finds look-alike characters — Cyrillic а that reads as a, Greek ο that reads as o, fullwidth digits — and can replace them with their ASCII equivalents.
- No length limits
- No registration
- Free forever
- Your text never leaves your browser
How to detect confusable characters
Paste the suspect text
Usernames, domains, identifiers, codes copied from documents.
Read the findings
Each confusable appears with its code point and the ASCII character it imitates.
Decide the policy
For identifiers you control, replace; for content you display, annotate.
Fix in one click
Enable the replace option and copy the ASCII-safe output.
Why confusables are a security issue
They defeat human review: two strings that look identical can be different identifiers at the byte level.
They enable spoofing: brand names, package names and public figures’ handles are the usual targets.
They slip past naive validation: length checks and regexes pass while the string means something else.
Confusable families covered
| Cyrillic | а с е о р х у and capitals А В С Е О Р |
|---|---|
| Greek | ο ν and capitals that mirror Latin shapes |
| Fullwidth | ０-９ and letters from CJK compatibility forms |
| Mathematical | 𝟎-𝟗 digit styles from the math alphanumerics |
| Action | Report only, or report + ASCII replacement |
Your text never leaves your browser
Every tool runs locally on your device. Nothing you paste is uploaded, stored, or tracked.
Frequently asked questions
Is “аpple.com” with a Cyrillic а really dangerous?
Yes — it is a different domain. This is the classic homoglyph attack, and the reason IDN rules forbid mixing scripts in most labels.
Does it catch every confusable?
It covers the high-frequency families (Cyrillic, Greek, fullwidth, math digits). The full Unicode confusables list has thousands of entries; this tool targets the ones that appear in real attacks.
Will replacing confusables break legitimate text?
Only if you feed it intentional foreign-language text — for identifiers, codes and usernames the ASCII replacement is exactly what you want.