Browser Guard
Monitor and govern AI usage across 500+ platforms directly in the browser — no proxy, no TLS interception.
The Governance Shield is a browser extension for Chrome, Edge, and Firefox that intercepts AI interactions at the DOM level before they leave the browser. It provides two layers of protection: broad visibility across 500+ AI platforms (Shadow AI Discovery) and deep prompt-level governance on 8 major LLMs.
How It Works
The extension operates entirely within the browser — no network proxy, no TLS break-and-inspect, no VPN. Prompts are evaluated against your gateway policies before they are sent to the AI platform.
User types prompt → Extension intercepts → Gateway policy check → Decision
↓
ALLOW → Prompt sent normally
BLOCK → Prompt not sent, user notified
WARN → User may proceed (logged)
JUSTIFY → User must provide reason (logged)
REDIRECT → User directed to approved toolTwo-Tier Detection
| Tier | Coverage | Capability |
|---|---|---|
| Tier 1 — Shadow AI Discovery | 500+ AI domains | URL-based detection, usage tracking, category classification |
| Tier 2 — Deep Governance | 8 LLM platforms | Prompt interception, PII detection, policy enforcement, response monitoring |
Tier 2 Platform Support
| Platform | Method | Enforcement |
|---|---|---|
| ChatGPT | Fetch intercept | Block, Warn, Justify, Redirect, Modify |
| Claude | Fetch intercept | Block, Warn, Justify, Redirect, Modify |
| Microsoft Copilot | WebSocket intercept | Block, Warn, Justify, Redirect, Modify |
| Perplexity | SSE intercept | Block, Warn, Justify, Redirect, Modify |
| Grok | Fetch intercept | Block, Warn, Justify, Redirect, Modify |
| DeepSeek | XHR intercept | Block, Warn, Justify, Redirect, Modify |
| Gemini | XHR intercept | Block only (Protobuf) |
| Meta AI | Fetch intercept | Block, Warn, Justify, Redirect, Modify |
Enforcement Actions
Each policy can define a browser enforcement action:
| Action | User Experience | Trace Record |
|---|---|---|
| BLOCK | Prompt is not sent. Red toast notification. | decision: BLOCKED |
| WARN | Amber warning. User may click "Send Anyway". | decision: WARNED, userAction: PROCEEDED |
| JUSTIFY | Dialog requires a written justification before proceeding. | decision: JUSTIFIED, justification: "..." |
| REDIRECT | Toast with link to an approved alternative tool. | decision: REDIRECTED |
| MODIFY | PII is masked automatically before sending. Green toast. | decision: MODIFIED, redactions: [...] |
Availability
Browser Guard is available starting from the Business tier.