NAZAR — Local Privacy Firewall

NAZAR — Let the agent see the page, not your secrets. Your browser is becoming the new interface for AI agents. But when an agent sees your screen, it may also see your passwords, emails, faces, IDs, financial details, and other private information. NAZAR puts a privacy gate between your browser and AI. It runs locally in your browser and uses DOM analysis, OCR, and vision to understand what is on your screen and identify sensitive information before it reaches an AI agent. 🔐 PRIVACY BEFORE AI Sensitive information is detected and masked locally before context is shared. Your raw page does not need to be exposed to the agent. 👁️ DOM + OCR + VISION Nazar doesn't rely on just one detection method. It combines webpage structure, text extraction and visual understanding to detect sensitive content across real webpages, images and visual elements. 🛡️ SEE WHAT THE AGENT SEES Preview the sanitized version before sharing context. Know exactly what is visible to the AI. 🤖 LET AI ACT — SAFELY Nazar doesn't replace your AI agent. It acts as a privacy gateway. The agent can understand the sanitized page and perform actions such as clicking, typing, scrolling and navigating. 🌐 YOUR CHOICE OF AI Nazar is designed to work as a privacy layer between your browser and the AI agent you choose — rather than locking you into one model or provider. ⚡ BUILT FOR YOUR DEVICE Choose lightweight models for resource-constrained devices or more capable models when your hardware supports them. ✋ MANUAL REDACTION Want complete control? Select any region and hide it yourself. 🖼️ SANITIZE IMAGES TOO Have an image you want to use with ChatGPT or another AI? Upload it to Nazar, generate a sanitized version, and use the protected image wherever you need. 🌍 WORKS ACROSS THE WEB Designed for 15+ modern browsers and real-world websites. The idea is simple: AI should understand what you want it to do — without needing to know everything about you. NAZAR Let the agent see the page, not your secrets.