AgentCloak

Sovereign AI that protects your private data in AI chatbots. Private data is cloaked locally and substituted before it is sent to AI. AgentCloak protects your private data before it ever reaches an AI chatbot. Everything happens on your device - no private data leaves your browser. HOW IT WORKS • As you type a prompt, AgentCloak detects private data directly in your browser with a small local AI. • Names, emails, phone numbers, SSNs, and more are replaced with cloaked substitute values before the message is sent to the AI. • When the AI responds, your AgentCloak digital twin restores your real data in the AI reply. PRIVACY BY DESIGN • 100% on-device processing. Your sensitive data is never sent to AgentCloak or any server. • The AI provider only ever sees the cloaked placeholders, not your real data. • No account, no login, no tracking. SUPPORTED CHATBOTS ChatGPT, Claude, Gemini, Grok, and DeepSeek. SUPPORTED LANGUAGES English, Spanish, French, German, Italian, Portuguese, and Dutch. Names in non-Latin scripts (e.g. Chinese, Japanese, Korean, Arabic, Cyrillic) are out of scope. SUPPORTED SENSITIVE DATA TYPES Names, phone numbers, email addresses, SSNs, credit cards, bank account / routing / tax IDs, government IDs, passports, driver's licenses, street addresses, IP addresses, and URLs. HARM REDUCTION, NOT A GUARANTEE No on-device model catches 100% of sensitive data. AgentCloak is a strong extra layer of protection — always review sensitive messages before sending. Includes Rampart by National Design Studio, licensed under CC BY 4.0