Customization
Aporia Guardrails is highly customizable, and we continuously add more customization options. Learn how to customize guardrails for your needs.
Get Started
To begin customizing your project, enter the policies tab of your project by logging into the Aporia dashboard, selecting your project and clicking on the Policies tab.
![](https://mintlify.s3-us-west-1.amazonaws.com/aporia/images/policies-tab-customization.png)
Here, you can add new policies1, customize2, and delete existing ones3.
A policy in Aporia is a specific safeguard against a single LLM risk.
Examples include RAG hallucinations, Restricted topics, or Prompt Injection. Each policy allows for various customizations, such as adjustable sensitivity levels or topics to restrict.
Adding a policy
To add a new policy, click Add policy to enter the policy catalog:
![](https://mintlify.s3-us-west-1.amazonaws.com/aporia/images/policy-catalog.png)
Select the policies you’d like to add and click Add to project.
Editing a policy
Next to the new policy you want to edit, select the ellipses (…) menu and click Edit configuration.
![](https://mintlify.s3-us-west-1.amazonaws.com/aporia/images/edit-policy.png)
Overview of the edit configuration page:
- Policy Detection Customization: Use this section to customize the policy detection algorithm (e.g. topics to restrict). The configuration options here depend on the type of policy you are editing.
- Action Customization: Customize the actions taken when a violation is detected in this section.
- Sandbox: Test your policy configurations using the chatbot sandbox. Enable or disable a policy using the Policy State toggle.
- Save Changes: Click this button to save and implement your changes.
The Quickstart guide includes an end-to-end example of how to customize a policy.
Deleting a policy
To delete a policy:
- Log into your Aporia Guardrails account.
- Select your project and click on the Policies tab.
- Next to the policy you’d like to remove, select the ellipses (…) and then select Delete policy from the menu.
Custom policies
You can also build your own custom policies by writing a prompt. See the Custom Policy documentation for more information.