User Guide

​

Step 1: Register Model (Mandatory)

NOTE: Currently, Kentron only supports OpenAI

​

Step 2: Create an Application

A central container for managing and monitoring each application’s security policies and guardrails. Every application starts with a default policy called “Launch,” which you can modify or remove as needed. The “Launch” policy consists of the following advance guardrails

• Toxicity Detection
• PII Detection
• Jailbreak Detection
• Profenitiy Detection

​

Step 3: There are three ways you can enforce the security policy in your agentic or AI/ML application

• Python
• JavaScript
• CURL

​

Step 4: The following are the required parameters to integrate Kentron with your applications:

• Polic ID: Gets auto-created
• Model ID: gpt-4o (case sensitive)
• Kentron API Key

 

​

Step 5: Playground

Test your existing or newly created policies in the Playground with ease, powered by built-in prompts designed to accelerate evaluation and refinement.