---
title: "Configuring Advanced Agent Settings"
description: "Introduction"
canonical_url: "https://help.brainfi.sh/articles/configuring-advanced-agent-settings-o5iQRqDdRS"
md_url: "https://help.brainfi.sh/articles/configuring-advanced-agent-settings-o5iQRqDdRS.md"
---
# Configuring Advanced Agent Settings

## Introduction
This guide covers the advanced configuration options available for fine-tuning the agent's technical behavior and integration. These settings are intended for administrators and are located in the "Advanced Settings (Super Admin)" section.

## Steps
1. From the agent configuration page, scroll down and expand the **Advanced Settings (Super Admin)** section.
2. Optionally assign a **CSS class name** for custom styling of the embedded agent.
3. Choose the **Agent version** from the dropdown menu to control which runtime or feature set the agent uses.
4. Toggle various behaviors such as:
   * **Enable tracking** – control whether analytics and user interaction tracking are enabled for this agent.
   * **Hide answers** – suppress displaying generated answers in the UI (useful for testing or special workflows).
   * **Hide feedback** – remove thumbs up/down or similar feedback controls from the agent interface.
   * **Enable follow-up questions** – allow the agent to prompt users with additional suggested questions.
5. Customize the instructional/system text for different states, such as:
   * Loading text while the agent initializes.
   * Text shown while an answer is being generated.
   * Any other status or helper messages exposed in this section.
6. Adjust the **Temperature** setting to control the randomness/creativity of the AI's responses. Lower values result in more focused and deterministic answers; higher values can produce more varied or exploratory responses.
7. At the bottom of this section, view the entire agent configuration as a raw **JSON config** object. This can be copied for debugging, backup, or programmatic configuration purposes.

## Summary
The Advanced Settings panel provides administrators with deep control over the agent's technical operations and presentation. Key capabilities include:

* Assigning CSS class names for targeted styling.
* Selecting the appropriate agent version.
* Toggling tracking, answer visibility, feedback collection, and follow-up questions.
* Customizing system and status messages.
* Controlling response creativity via the Temperature setting.
* Inspecting the full agent configuration in JSON format for advanced debugging or integration workflows.