Clean up text · homoglyphsCyrillic letters in English text
A word looks English, but a Cyrillic “а” or “о” is hiding inside — and search, spell-check and string comparison no longer recognise it. We find such words and fix them.
The text is processed in your browser — nothing is sent to a server
foreign letters are underlined — hover to see the code
Input · source text
Your аccount pаssword expires today — log in at the official pаge to keep access. Search could not find the word “tаble” in the report.
Мoscow is mostly Latin, so it is fixed. Mosкwа has a letter with no pair and is only flagged.
A tie stays as it is: Тоxy.
Your аccount pаssword expires today — log in at the official pаge to keep access. Search could not find the word “tаble” in the report.
Мoscow is mostly Latin, so it is fixed. Mosкwа has a letter with no pair and is only flagged.
A tie stays as it is: Тоxy.
аccount→accountmostly Latin — Cyrillic letters replaced
pаssword→passwordmostly Latin — Cyrillic letters replaced
pаge→pagemostly Latin — Cyrillic letters replaced
tаble→tablemostly Latin — Cyrillic letters replaced
Мoscow→Moscowmostly Latin — Cyrillic letters replaced
Mosкwа→not fixeda letter has no twin — only flagged
Тоxy→not fixeda tie — left as is
03
About homoglyphs
More than a dozen letters look the same in the Latin and Cyrillic alphabets. Mixed words appear when copying from PDFs, when typing with the wrong keyboard layout, and on purpose — to slip past filters or fake a brand name. We check every word: if it has letters from both alphabets, we highlight them and offer to replace them with letters of the alphabet that dominates the word.
Lookalikes
18 pairs
a а · o о · c с
Lowercase a, e, o, p, c, y, x and uppercase A, B, E, K, M, H, O, P, C, T, X match Cyrillic letters by shape.
Why it matters
Search
0 matches
The word “table” with a Cyrillic “а” is found neither by site search, nor by Ctrl+F, nor by a database query.
Where they come from
PDF and layout
Pаssword
Most often from copying out of PDFs and typing with the wrong layout. Less often — on purpose, in phishing.
Filters
Bypass
will not work
Spam filters and plagiarism checkers map lookalikes to one alphabet before comparing, and flag many of them as suspicious.
How we fix
By majority
Pаssword → Password
If a word has more Latin letters, the Cyrillic lookalikes become Latin, and the other way round.
Standard
UTS #39
confusables
Lookalike characters are described in the Unicode security standard — browsers use it to defend against fake domains.
04
Frequently asked questions
Paste the text into the box: words that mix Latin and Cyrillic letters are highlighted, and the Cyrillic letters in them are underlined. Hover over a letter to see its code point — U+0430 for the Cyrillic “а” against U+0061 for the Latin “a”. Below the table of pairs you get a list of every mixed word: how it will be fixed or why it is better left alone.
Lowercase а, е, о, р, с, у, х and uppercase А, В, Е, К, М, Н, О, Р, С, Т, Х — eighteen pairs that are indistinguishable in most fonts. In some typefaces other pairs look alike too, such as lowercase к and k or the italic т and m. The tool only highlights those and does not fix them: replacing a letter on looks alone is too risky.
Turn on “Fix by majority”: in each mixed word the lookalikes are replaced with letters of the alphabet that dominates the word. “Pаssword” with a Cyrillic “а” becomes “Password”, and “Mосква” becomes “Москва”. If a word has as many letters of both alphabets, or a minority letter has no twin, the word stays as it is and is flagged in the list.
To make a fake look like the real thing. A domain or brand name typed with a Cyrillic “а” or “о” looks identical on screen but is a different string, so it can pass filters and fool a reader. Browsers defend against such domains using the Unicode security standard. Pasting a suspicious name here shows at once which letters are not what they seem.
Most likely the word contains a Cyrillic lookalike. To a computer “table” with a Cyrillic “а” is a different word: the letter has another code, so neither Ctrl+F, nor site search, nor a database query will find it. Paste the fragment here — the word is highlighted, and fixing it brings back plain Latin letters, after which search works again.
Most often from PDFs and from typing with the wrong keyboard layout switched on for part of a word. Less often lookalikes are inserted on purpose: to slip past spam and moderation filters or to pass off a fake web address as a real one. Checking is worth it either way: such text looks perfectly normal, and the errors surface later — in search and databases.