How SENTINEL Works
Two ways to protect your AI interactions — explained without the technical jargon.
Path 1: Chrome Extension
Real-time protection while you browse. No code, no configuration, no hassle.
Think of It Like a Security Camera for Your AI Conversations
You install a security camera in your shop — it watches everything, alerts you when something suspicious happens, and keeps a record. SENTINEL's Chrome Extension does the same for every AI conversation in your browser. It watches what goes in and what comes out, scores both for signs of jailbreaking or prompt injection, and alerts you the moment something looks wrong.
Get Your API Key
Sign up for a free SENTINEL account. You'll receive your API key by email — it takes less than a minute.
Install & Activate
Download the Chrome Extension, load it in your browser, and paste your API key. Then use ChatGPT, Claude, Gemini, Grok, DeepSeek, or Meta AI as you normally would. SENTINEL runs silently in the background, scoring every message.
Get Alerted When It Matters
If a jailbreak attempt or suspicious prompt is detected, SENTINEL shows a Chrome notification immediately. Your shield icon changes color so you can see the status at a glance — green means clear, yellow means caution, red means a threat was found.
Dual Scanning
Scores both your input and the AI's response. A jailbreak prompt is caught even before the AI responds.
Per-Tab Tracking
Every tab has its own security status. Switch between AI platforms and each conversation is monitored independently.
Exportable History
Download your scan history as JSON — timestamps, scores, platforms, and sources for auditing or compliance records.
Path 2: Compliance API
Tamper-proof certificates for AI agents you build and deploy.
Think of It Like a Notarized Audit
A bank gets audited to prove it handles money correctly. An accountant reviews the books, stamps a certificate, and anyone can check that certificate. The bank doesn't have to show its customer data — just prove it followed the rules. SENTINEL does the same for AI agents. It generates a tamper-proof certificate proving behavior was within your defined limits — and anyone can verify that certificate without seeing the underlying data.
Set Your Rules
Define what "good behavior" means for your AI agent. These are the boundaries — quality scores, response appropriateness, or safety thresholds. You set the limits, SENTINEL monitors them.
SENTINEL Certifies
SENTINEL continuously watches your agent's behavior. When behavior stays within limits, it generates a compliance certificate automatically — in the background, with no action needed.
Anyone Can Verify
Share your certificates with regulators, customers, or auditors. They can confirm the certificate independently — no SENTINEL account needed, no data exposed.
What Makes SENTINEL Different
Works at Both Ends
Browser extension for people who use AI. Compliance API for people who build AI. Same behavioral intelligence engine underneath, two different interfaces.
Tamper-Proof
Compliance certificates use advanced mathematics to ensure they can't be forged, altered, or backdated. If someone tampers with a certificate, verification fails instantly.
Privacy-Preserving
Your certificates prove compliance without revealing actual data. The extension scores behavior locally. The API generates proofs that prove "within limits" — never individual observations.
Your Compliance Timeline
Each certificate links to the previous one, forming a chain. Any gap or modification is immediately detectable.
Certificate #1
January
Certificate #2
February
Certificate #3
March
Every certificate proves its predecessor is genuine. No gaps. No modifications. An unbroken chain of continuous compliance.
Ready to Get Started?
Get a free API key, install the Chrome Extension, and start protecting your AI conversations today.