Detection
Regex-based PII scanner covering 15+ API key formats, branded credit cards, international phone numbers, emails, and more—all evaluated client-side in real time.
Security Layer
Sentinel
A high-fidelity, zero-knowledge security layer for AI LLM interfaces. Real-time PII detection, automated masking, and an encrypted prompt library—all running locally in your browser, never touching our servers.
Zero-Knowledge
No server, no data
Real-time PII
Regex-based scanning
AES-256-GCM
Encrypted prompt vault
Masking
"Mask All" or selective masking lets you strip sensitive content before it reaches the AI. You decide the balance between privacy and prompt utility.
Prompt Library
Save, organise, and export your favourite prompts. Each entry is encrypted with AES-256-GCM via the Web Crypto API using PBKDF2-derived keys—only your extension context can decrypt them.
Architecture
Built as a client-side IIFE injected via Chrome content scripts. Detection runs in the browser's memory with zero cloud calls. A secure message bus between service worker and components prevents cross-extension sniffing.
Permissions
Least-privilege model—storage, activeTab, and scripting permissions scoped to ChatGPT, Claude, and Gemini only. No broad "access all websites" requests, no trackers, no analytics.
Sentinel Privacy Policy
Understand how Sentinel's zero-knowledge architecture ensures your data never leaves your browser.
Read privacy policySentinel Terms of Use
Review the agreement governing your use of Sentinel, including license, liability, and accuracy disclaimers.
Read terms of use