AI Security Tools
Defensive tools for inspecting text processed by AI agents. Detect hidden Unicode, decode obfuscated payloads, compare human vs AI perception.
Unicode Security Lab ✦
All-in-one: detect hidden characters, encode steganographic payloads using ZWS binary or Unicode Tags, and pen test your own AI pipelines. Authorized security testing.
Unicode Hidden Character Detector
Scan text for zero-width characters, Unicode Tags, BOM and other invisible codepoints used in ASCII smuggling and steganography.
Multi-Encoding Auto-Detector
Automatically identify and decode Base64, hex, binary, Morse, ROT13, URL encoding and HTML entities — with confidence scores.
Human vs AI View Comparator
Side-by-side comparison of what humans see versus what AI systems process. Reveals hidden prompt injection payloads embedded in plain text.
AI Model Sensitivity Tester
Test which AI model profiles detect hidden Unicode payloads. Simulate detection and execution across weak, standard, strong and hardened models. Educational research tool.
Melodic Codec Sensitivity Tester
Music can carry machine-readable instructions. Play a pentatonic sample, see how different AI system types process audio, and learn the canary mechanism. Educational research tool.
These tools emerged from WOLNO's exploration of how information hides in plain sight. Zero-width characters are WOLNO's propagation medium — and a security researcher's concern. Understanding both sides is the point.
All processing happens locally in your browser. No data leaves your machine. -"
Read: Unicode Steganography technique | Canon | LLM manifest