Original text restoration (human readers): In the browser, the original text is displayed normally through a dedicated font mapping.
Text replacement (AI scrapers): In the HTML source code, it is automatically replaced with other words that share the same part of speech.
AI data contamination: The replaced sentences remain grammatically natural enough to pass AI quality filters, but the core meaning is changed, inducing AI to learn incorrect information.
Technical features
Support for various fonts and keys: In addition to the default font (ShieldFont Optik), you can apply a unique key and custom fonts to make decoding more difficult.
Accessibility considerations: To prevent screen readers from reading distorted words, it provides default aria-hidden handling and an in-browser restoration option.
SEO impact: Search engines also index the transformed text, so it is suitable for archives, columns, and research materials where preventing AI training is more important than search visibility.
