> ## Documentation Index
> Fetch the complete documentation index at: https://hydroxai.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Firewall

> Real-time input/output validation for AI applications.

The Know Your AI Firewall provides real-time validation of AI inputs and outputs, catching policy violations,
harmful content, and security risks before they reach your users.

## How it works

The firewall sits between your application and your AI model, validating each request and response:

1. Your application sends the user input through the firewall
2. The firewall validates the input against configured rules
3. The request is forwarded to your AI model
4. The model response is validated before being returned to the user
5. Each validation is logged with risk categorization and scoring

## Setup

<Steps>
  <Step title="Generate an API key">
    In your product page, go to **Firewall** and generate a firewall API key.
  </Step>

  <Step title="Integrate the SDK">
    Add the `@know-your-ai/firewall` SDK to your application and configure it with your API key.
  </Step>

  <Step title="Monitor validations">
    View all validations in **Firewall Logs** with risk details and pass/fail status.
  </Step>
</Steps>

## Firewall logs

The **Logs** page shows a complete history of all firewall validations with:

* **Risk category** — The type of risk detected (e.g., jailbreak, harmful content, PII)
* **Risk score** — Numeric confidence score for the detection
* **Risk reason** — Explanation of why the content was flagged
* **Pass/fail status** — Whether the validation passed or was blocked
* **Input/output tokens** — Token count tracking per validation

## Token usage analytics

The **Token Usage** page provides time-series charts for:

* Input tokens over time
* Output tokens over time
* Total token count
* Configurable granularity: hour, day, month
* Time ranges: 24 hours, 7 days, 30 days, 90 days

## Security report

The **Security Report** provides a visual before/after comparison:

* **Baseline results** — How your model performs without the firewall
* **Firewall results** — How your model performs with firewall protection
* Per-prompt pass/fail comparison table
* Animated security score counters
* Compliance panel showing regulatory impact

## Related docs

<CardGroup cols={2}>
  <Card title="Evaluation" icon="check-double" href="/evaluation">
    Run security evaluations to test your model.
  </Card>

  <Card title="Monitoring" icon="chart-line" href="/monitoring">
    Track production usage with SDK monitoring.
  </Card>
</CardGroup>