Playground

Test AI responses with different models and settings.

The Playground APIs let you test your chatbot's AI responses with different configurations — try different models, prompts, and temperatures without changing your project's saved settings.

How It Works

  1. Create a session — call Create Playground Session to get a session_uid
  2. Test responses — call Get Playground Response with different agent_setting overrides
  3. Compare — change model, prompt, or temperature and compare the AI outputs
  4. Apply — once you find the best configuration, save it using Update Project Settings

Permissions

EndpointRequired permission
Create Playground SessionConversation > Write
Get Playground ResponseDebugging > Read

On this page