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

# Agents

> Create and manage your AI agent profiles

# Agents

Create and configure your AI applications. Each agent profile represents a distinct AI application with its own configuration, guardrails, and monitoring.

## Creating an Agent

<Steps>
  <Step title="Navigate to Agents">
    Click **"Agents"** in the sidebar
  </Step>

  <Step title="Create New Agent">
    Click **"Create Agent Profile"** → Fill in:

    * **Agent Name**: Descriptive name (e.g., "Customer Support Bot")
    * **Agent Purpose**: Purpose and behavior guidelines
  </Step>

  <Step title="Assign Guardrails">
    Enable gaurdrails applicable to this Agent. Halios will customize each gaurdrail based on Agent's persona you defined earlier.
  </Step>

  <Step title="Save Agent">
    Get your unique **Agent ID** for API integration
  </Step>
</Steps>

## Agent Management

* You can review guardrails associated with Agent.
* You can assign new guardrails. Only guardrails currently unassigend to any other Agent are available to assign. You can also create new guardrail from template.
* You can enabled/disable guardrail evaluation for request and responses.

## Zero Data Retention

* You can enable zero data retention policy for an agent. Halios won't store, log raw request/responses if this policy is ‘enabled’.

## Best Practices

* Describe agent's purpose thoughtfully.
* Consider enabling prompt injection and persona integrity guardrail for each agent.
* Sensitive data guardrail introduces latency, as Halios will evaluate sensitive data guardrail first and only used redacted messages for evaluating other guardrails.
* Use different agents for dev/test/prod workload and testing.

***

**Next Steps**: [Configure Guardrails](/dashboard/guardrails) for your agent or explore the [API Playground](/dashboard/playground) for testing.