Dashboard

Real-time monitoring of AI gateway activity

Enterprise AI Security

Protecting sensitive data from external AI exposure

0
Total Scanned
0
Protected
0
Detections
AgentShield Enterprise
0
Total Requests
0
Allowed to External
0
Blocked
0
Routed to Local LLM
Configuration
Test Chat
View Logs

Recent Activity

No recent activity

Configuration

Configure detection patterns, roles, integrations, and scale settings

Sensitive Patterns
Roles & Access
AD Integration
SIEM / QRadar
Scale & Quotas

Sensitive Data Patterns

Define regex patterns to detect sensitive data. These patterns will trigger blocking or local LLM routing. Standard patterns for PAN, Aadhaar, and API keys are pre-configured.

Name Regex Pattern Severity Actions

Organization-Specific Patterns

Regex pattern for your client codes

Regex pattern for account identifiers

Regex pattern for internal identifiers

Regex pattern for unique codes

Regex pattern for employee IDs

Regex pattern for project codes

Sensitive Keywords

Enter keywords that indicate sensitive information. One keyword per line.

Role-Based Access Control

Define roles with specific permissions for accessing sensitive data and AI capabilities. Map these roles to AD groups for automatic provisioning.

Role Name Permissions AD Group Actions

Active Directory Integration

SIEM Integration

IBM QRadar
Not Connected
Splunk
Not Connected
Elastic SIEM
Not Connected

Scale Configuration

Test Chat

Test the AI gateway with sample prompts

Interactive Chat

Start a conversation to test the gateway

Scan Tester

Test sensitive data detection patterns

Input Text

Scan Results

Results will appear here

Request Logs

View all gateway requests and their status

Timestamp User Status Detections Latency
No logs available

Detection Patterns

Manage sensitive data detection patterns

Active Patterns

Name Regex Pattern Severity Actions

Usage & Telemetry

Monitor system usage and performance metrics

99.9%
Uptime
42ms
Avg Latency
2.4GB
Memory Usage
23%
CPU Usage

Agent Registry

Manage registered AI agents and their configurations

No agents registered yet

MCP Gateway

Model Context Protocol - Enterprise Tool Access Control

-
Registered Tools
-
Active Tools
-
Pending Approvals
-
Access Requests (24h)

Enterprise Tools Registry

MCP Gateway controls which AI agents can access enterprise tools. Tools with UPSI or RESTRICTED classification require explicit approval for each access request.

Loading enterprise tools...

Access Audit Log

No access logs yet

Secrets / KMS

Enterprise Key Management System - Securely manage API keys and secrets

KMS Provider Configuration

Checking...

Enterprise KMS: Choose where your secrets are stored. For development, use Local Vault. For production, integrate with AWS Secrets Manager, Azure Key Vault, or HashiCorp Vault.

Loading...
Checking connection...

Local Vault uses AES-256 encryption. Suitable for development and small deployments. For high availability, use cloud KMS.

For production, set via VAULT_MASTER_KEY environment variable

Secrets Management

No secrets configured

Custom AI Providers

Add custom AI providers (Sarvam, Moonshot, etc). Once added, they appear in all dropdowns automatically.

No custom providers configured

Red Team Scanner

AI security testing and adversarial simulation

Attack Simulation

Red team testing simulates 50+ attack vectors including prompt injection, jailbreaks, data exfiltration, and more. Tests run against your AI Gateway's chat endpoint.

The AI endpoint to test (auto-filled with gateway chat endpoint)
✓ = API key configured

AI Provider Settings

Configure model routing and local LLM for sensitive data

Dual-Routing Architecture: Clean data routes to cloud providers. Sensitive data (PII, confidential) routes to your Local LLM, keeping it within your network.

Local LLM (On-Premise)

For Sensitive Data

Requests containing PII, financial data, or confidential information are automatically routed here. Data never leaves your network.

Cloud Providers

For Non-Sensitive Data

Cloud provider API keys are managed in the Secrets/KMS section. Configured providers appear in the Chat interface automatically.

Loading providers...