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

# Airia Guardrails: Secure AI Made Simple

<Card title="Build AI Applications with Confidence" icon="sparkles">
  Transform your AI development with Airia Guardrails – the intelligent safety net that lets you innovate fast while staying secure. Whether you're building customer-facing chatbots or internal AI tools, our configurable safeguards work seamlessly with any foundation model to protect your business and users.

  **Why Airia Guardrails?**

  * **Universal Protection**: Works with any AI model or framework
  * **Flexible Control**: Apply safeguards organization-wide, by project, or to specific agents
  * **Comprehensive Coverage**: Protects against security threats, inappropriate content, and data breaches
  * **Smart Filtering**: Monitors both what goes into your AI and what comes out
</Card>

<Steps>
  <Step title="Create a Guardrail">
    * Go to **Security Tab**
    * Select **Runtime Security** from the left navigation
    * Click **Guardrails** this will open the existing guardrails (in any) or click **Create Guardrail** to create one by giving it a name and description
  </Step>

  <Step title="Filters">
    * Click [**Filters**](/admin-hub/policies/filters) - discover the different filter and protection options
    * Choose your filters – mix and match as needed
  </Step>

  <Step title="Target Your Protection">
    * Select**Target and Scope**  to define coverage
    * Choose what to protect: AI Gateway traffic, Model inputs, outputs, or both
    * Set your scope: All agents or specific agents per project
    * **Save draft** to test or **Apply Guardrail** to activate
  </Step>
</Steps>

<Tip>
  **Getting Started**: We recommend testing guardrails with one agent or project first. This allows you to validate your filter settings and understand their impact before applying them organization-wide.
</Tip>