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

# Settings

> Configure your AI Agent behavior and preferences

The Settings page allows you to configure your AI Agent's behavior, name, and operational modes. Settings are saved automatically and applied immediately.

## Agent Information

### Agent Name

**Edit your agent name**

* Click the agent name to edit it inline
* Changes are saved automatically
* The name appears in the agent dashboard and can be used for identification

**Last Updated**

* Shows when the agent was last modified
* Updates automatically when you make changes to any setting

## AI Mode

Control how your agent handles incoming conversations.

### Mode Options

**AI Enabled** (Default)

* Agent automatically responds to customer queries using AI
* Uses training data to provide relevant answers
* Responds in real-time without human intervention
* Best for: High-volume customer support, FAQ handling, 24/7 availability

**Manual Mode**

* Disables automatic AI responses
* All conversations require human agent assignment
* Useful for: Sensitive inquiries, complex cases, testing, or when AI needs supervision

### How to Toggle Mode

1. Navigate to Settings
2. Find the **AI Mode** toggle
3. Switch between **Enabled** and **Disabled**
4. Changes apply immediately to new conversations

<Note>
  Existing conversations maintain their current mode. The toggle affects new incoming conversations.
</Note>

## Off-Topic Guardrails

Control whether your AI Agent should stay focused on topics related to your business or answer any question.

### Guardrail Options

**Stay On Topic** (Recommended)

* Agent only answers questions related to your training data
* Politely declines off-topic queries
* Keeps conversations focused on your business
* Best for: Customer support, product inquiries, business FAQs

**Open**

* Agent attempts to answer any question
* Uses general knowledge beyond training data
* More flexible but may stray from business topics
* Best for: General assistants, flexible chatbots, exploratory use cases

### How to Configure

1. Navigate to Settings
2. Find the **Off-Topic Guardrails** toggle
3. Switch between **Stay On Topic** and **Open**
4. Changes apply immediately

<Warning>
  When set to "Open," your agent may answer questions unrelated to your business, which could use more credits and may not align with your support goals.
</Warning>

## Quick Actions

### Train Agent

* **What it does**: Opens the Training overview page where you can add or manage training data
* **When to use**: After gathering new content, creating FAQs, or when you want to update your agent's knowledge
* **Click**: "Train Agent" button redirects to Sources section

### Delete Agent

<Warning>
  **Danger Zone**: Deleting your agent is permanent and cannot be undone.
</Warning>

**What gets deleted:**

* Agent configuration and settings
* All training data (files, text content, Q\&A pairs, website sources)
* All conversations and message history
* Deployment configurations
* Analytics data

**Before deleting:**

1. Ensure you have backups of any important data
2. Remove the agent from your website if deployed
3. Notify team members if the agent is shared

**To delete an agent:**

1. Scroll to the bottom of Settings
2. Click **Delete Agent**
3. Confirm deletion in the dialog
4. Agent is permanently removed

## Real-Time Updates

Settings changes are applied immediately without page refresh. The system uses WebSocket connections to sync updates across:

* Multiple browser tabs
* Team member sessions
* Agent dashboard views

When a setting changes, all connected sessions receive the update automatically.

## Settings Summary

| Setting                  | Options              | Default       | Impact                        |
| ------------------------ | -------------------- | ------------- | ----------------------------- |
| **Agent Name**           | Custom text          | "New Agent"   | Display name in dashboard     |
| **AI Mode**              | Enabled / Manual     | Enabled       | How conversations are handled |
| **Off-Topic Guardrails** | Stay On Topic / Open | Stay On Topic | Topic restriction             |

## Best Practices

**Agent Name**

* Use descriptive names like "Customer Support Bot" or "Product Expert"
* Include purpose or department if you have multiple agents
* Keep names concise and clear

**AI Mode**

* Start with AI Enabled for most use cases
* Use Manual Mode temporarily during training or testing
* Monitor conversations to ensure AI quality before full automation

**Off-Topic Guardrails**

* Keep "Stay On Topic" enabled for customer-facing agents
* Use "Open" only for internal tools or exploratory purposes
* Review conversation logs to see if guardrails are working effectively

***

## What's Next?

<CardGroup cols={2}>
  <Card title="Sources" icon="database" href="/user-guides/agent/sources">
    Add training data to your agent
  </Card>

  <Card title="Activity" icon="message" href="/user-guides/agent/activity">
    View conversations and responses
  </Card>

  <Card title="Analytics" icon="chart-line" href="/user-guides/agent/analytics">
    Monitor agent performance
  </Card>

  <Card title="Test Agent" icon="flask" href="/user-guides/quick-start/your-first-agent#step-2-test--optimize-your-ai-agent">
    Test your agent's responses
  </Card>
</CardGroup>