Homoglyph Detector
Scan text or a domain for homoglyph attacks: Cyrillic and Greek characters that render identically to Latin letters (а vs a, р vs p). Returns every impostor position with its Latin equivalent and the fully normalized ASCII rendering — the definitive check for spoofed brands and IDN domains.
Worked examples
Spoofed 'paypal' (Cyrillic а, р)
GET /api/v1/security/homoglyph-detect?text=%D1%80%D0%B0ypal.com
Result: 2 look-alike character(s) detected — reads as "paypal.com"
Spoofed 'google' (Cyrillic о)
GET /api/v1/security/homoglyph-detect?text=g%D0%BEogle.com
Result: 1 look-alike character(s) detected — reads as "google.com"
Clean ASCII
GET /api/v1/security/homoglyph-detect?text=github.com
Result: ✓ Pure ASCII — no homoglyphs possible
Machine API (x402)
$0.01 / callThis tool is also a JSON API for AI agents. Requests without payment receive 402 Payment Required plus instructions; agents pay USDC on Base via the x402 protocol — no accounts, no API keys.
GET /api/v1/security/homoglyph-detect?text=%D1%80%D0%B0ypal.com HTTP/1.1
Host: agentstoolshub.com
→ 402 (payment required, instructions in headers)
→ 200 (after X-PAYMENT header; JSON body below)
{
"tool": "security/homoglyph-detect",
"input": {"text":"раypal.com"},
"result": { "value": 2, "answer": "2 look-alike character(s) detected — reads as "paypal.com"" }
}Agent docs: /llms.txt · OpenAPI spec · integration guide
About this tool
Detects the exact characters behind look-alike spoofing: Cyrillic 'а' inside 'paypal', Greek 'ο' inside 'google'. Every impostor is reported by position, Unicode codepoint and its Latin equivalent, plus the normalized ASCII rendering of the whole string. Pure analysis, instant, no lookup.
Frequently asked questions
Why do these characters fool people?
Cyrillic а, е, о, р, с, х, у and their Greek cousins render pixel-identical to Latin letters in most fonts, but are different Unicode codepoints — so 'раypal.com' is a completely different domain that reads as paypal.com.
Which characters are covered?
The high-frequency confusables: 20+ Cyrillic and Greek letters with Latin twins, covering effectively all real-world IDN spoofing in abuse reports. Full Unicode confusables tables (UTS #39) are far larger but add little practical coverage.