Overview of Patronus AI

Patronus AI is a tool designed to help detect and correct mistakes in large language models (LLMs) at scale, ensuring that AI-generated content meets quality standards. By integrating Patronus AI with Naologic, you can monitor and refine generative AI outputs, allowing for confident use of AI within your application.

With the Patronus AI integration, you can:

• Monitor AI Outputs: Detect errors or inconsistencies in content generated by AI models.

• Ensure Content Quality: Use Patronus AI to improve the reliability and accuracy of generative AI within your application.

• Enhance Trust in AI: Scale generative AI confidently by catching and mitigating common LLM mistakes.

Installing the Patronus AI API in Naologic’s Builder

To set up the Patronus AI API for your project, follow these steps:

1. Install the Patronus AI API:

   • Go to your app's canvas.

   • Click on the plus sign on the left to open the element selection menu.

   • Navigate to the APIs section.

   • Search for "Patronus AI" and press Install to add it to your project.

Connecting Naologic to Your Patronus AI Account

To link Naologic with your Patronus AI account, follow these steps:

1. Access Integration Settings:

   • Navigate to the top right of the Workspace and click on the cog icon (Settings).

   • Select Integrations from the settings menu.

2. Connect to Patronus AI:

   • In the Integrations menu, find Patronus AI and click Connect.

   • Enter the following information from your Patronus AI account:

     • API Key: Your Patronus AI API Key.

   • Once the API Key is entered, click Connect to establish the connection between Naologic and your Patronus AI account.

Using Patronus AI within Your Application

Since Patronus AI does not have a dedicated action within Naologic, it can be integrated in the background or as part of existing workflows. For custom configurations or complex workflows that integrate Patronus AI, consider reaching out to our support team for assistance in implementing specific use cases.