Protect your brand from lookalike domain attacks.
Spoofing isn't limited to Latin. d0ma1n scans across all 12 ICANN-approved IDN scripts.
Not a static homoglyph table. Every confusable pair is SSIM-scored across 230 fonts, with p95 and max metrics. 793 pairs are novel discoveries beyond TR39.
Latin, Cyrillic, Greek, Arabic, Han, Hangul, Katakana, Hiragana, Devanagari, Thai, Georgian, Armenian. Most tools only check Latin lookalikes.
Visual similarity depends on the font. A substitution that is invisible in Arial may be obvious in Georgia. d0ma1n finds the worst-case font for each variant.
Inspired by dnstwist. Built on measured visual similarity instead of static tables.
| Capability | |
|---|---|
| Confusable pairs | 1,397 SSIM-scored (793 novel, beyond TR39) |
| Scripts | 12 ICANN-approved IDN scripts, bidirectional |
| Scoring | Continuous 0 to 1 (p95 and max SSIM) |
| Font awareness | Best-font lookup across 74 fonts |
| DNS resolution | A, AAAA, MX, NS with threat classification |
| IDNA filtering | Only PVALID characters (registrable domains) |
| Reverse lookup | Identify what a suspicious domain impersonates |
| Runtime dependencies | 1 (namespace-guard) |
Three open-source projects in a pipeline.
Renders 230 fonts, computes SSIM similarity for every Unicode pair. Produces scored confusable maps.
Ships the maps as runtime data. Provides skeleton(), areConfusable(), and cross-script detection.
Inverts the maps, generates domain variants via k-edit enumeration, scores them, resolves DNS.
Built in the open. MIT licensed.