UPDATE (March 1, 2026): BCorrect has successfully transitioned to ShoyHuman 04.0hF04. Older framework versions (03.5h, 03.5hF02, etc.) mentioned in this document are now deprecated in favor of the full English Singularity (4.0 Series).
Updated the Free Scanner
Free scanner: /free-tools/invisible-char-checker
It's still just a basic online demo limited compared to what full binary-level scanning can do. But it's better than most free tools despite the constraints.
What Changed:
86 threats → 205+ threats (138% increase)
5 Cyrillic chars → 60+ confusables across 4 categories
What It Detects
| Category |
Count |
Examples |
| Cyrillic |
25+ |
pаypal.com, admіn, 2О24 |
| Greek |
15+ |
Αpple, Βank, Ιntel |
| Math Bold |
10+ |
ð€BC, ðŸŽ123 |
| Fullwidth |
10+ |
ＡＢＣＤ, ０１２３ |
Attack types:
- Brand spoofing (pаypal vs paypal)
- Variable injection (admÑ–n vs admin)
- Numeric substitution (2О24 vs 2024)
- Mixed evasion
How It Works
Confusability Score: 0.85-0.99
Critical Threshold: >0.95
Processing: Client-side, <100ms
False Positives: <2%
Test results:
- 47/500 brand spoofs caught
- 23 variable attacks in 10K lines
- 156/5K phishing URLs flagged
Limitations
This is character-level detection. It catches visible homoglyphs but misses:
- Encoding corruption (malformed UTF-8)
- Trojan Source attacks (CVE-2021-42574)
- Steganography in whitespace
- Byte-level exploits
Character filtering happens after corruption occurs. Binary scanning prevents it.
That's what the commercial system does operates at the byte level before interpretation.
Try It
Free scanner: /free-tools/invisible-char-checker
Advanced view: /tools/advanced-scanner
Enterprise (binary-level): /bad-character-scanner
It's free. It's limited. It works.
→ Try the free scanner now â†