Why Monitoring & Firewall?

Protect your AI applications in production — real-time monitoring of agent behavior and automated blocking of harmful content.

AI applications in production face two critical risks:

1. Unpredictable agent behavior — AI agents can hallucinate, leak sensitive data, call unauthorized tools, or drift from their intended goals
2. Harmful content generation — Models can produce toxic, biased, or policy-violating outputs that reach your end users

Know Your AI’s Monitoring and Firewall work together to address both:

Monitoring

See everything. Track every AI request, response, tool call, token, cost, and error in real time. Trace multi-step agent workflows as span trees. Catch anomalies before they become incidents.

Firewall

Block threats. Validate every input and output against content safety policies. Block jailbreak attempts, prompt injection, PII leakage, and harmful content — before they reach your users.

Architecture

User Input
    │
    ▼
┌─────────────────────────┐
│   Firewall (Input)      │  ← Block jailbreaks, prompt injection, PII
│   beforeRequest hook    │
└────────────┬────────────┘
             │  ✔ Safe
             ▼
┌─────────────────────────┐
│   AI Model              │  ← Your LLM (Gemini, GPT, Claude, etc.)
│                         │
└────────────┬────────────┘
             │
             ▼
┌─────────────────────────┐
│   Firewall (Output)     │  ← Flag harmful, biased, or toxic responses
│   afterResponse hook    │
└────────────┬────────────┘
             │  ✔ Safe
             ▼
┌─────────────────────────┐
│   Monitoring            │  ← Capture tokens, cost, latency, traces
│   Know Your AI Backend  │
└─────────────────────────┘
             │
             ▼
        User Response

What you can protect against

Threat	Monitoring detects	Firewall blocks
Jailbreak attempts	Logs the input for review	Blocks before reaching model
Prompt injection	Tracks anomalous inputs	Blocks injected instructions
PII leakage	Flags responses containing personal data	Blocks PII from being returned
Harmful content	Captures and categorizes output	Blocks toxic/hateful responses
Excessive agency	Traces all tool calls and agent steps	—
Cost spikes	Alerts on abnormal token usage	—
Latency degradation	Tracks TTFB and response times	—
Model errors	Captures error types and rates	—

Getting started

1. Install the SDK

   npm install @know-your-ai/node @know-your-ai/firewall
   
2. Get your DSN and Firewall API key

   From the Know Your AI dashboard:

   • DSN: Settings → API Keys
   • Firewall API key: Product → Firewall → Generate Key
3. Initialize with both integrations

   import * as KnowYourAI from '@know-your-ai/node';
   import { firewallIntegration } from '@know-your-ai/firewall';
   
   KnowYourAI.init({
     dsn: process.env.KNOW_YOUR_AI_DSN!,
     integrations: [
       KnowYourAI.googleGenAIIntegration(),
       firewallIntegration({
         baseUrl: process.env.FIREWALL_URL!,
         apiKey: process.env.FIREWALL_API_KEY!,
         onInputViolation: 'block',
         onOutputViolation: 'log',
       }),
     ],
   });
   
4. View dashboards

   Open the Monitoring and Firewall Logs pages in your product dashboard to see real-time data.

Next steps

Real-time monitoring

Set up production monitoring with dashboards, tracing, and alerts.

Content firewall

Configure input/output validation and block harmful content.

Agent safety

Monitor and protect multi-step AI agent workflows.

Production recipes

Copy-paste recipes for common production setups.