← Back to changelog

September 30, 2025

Structured Output Support for Prompt Experiments

Hassieb Pakzad

Enforce JSON schema response formats in prompt experiments to ensure consistent, parseable outputs for evaluation and analysis.

You can now enforce structured output schemas in prompt experiments, ensuring LLM responses conform to specific JSON formats. This makes it easier to evaluate outputs consistently and process results programmatically.

What’s New

• Structured output toggle: Enable or disable structured output for any experiment run with a simple toggle switch
• Schema reusability: Use the same schemas you’ve created in the Playground across all your experiments
• Easy schema management: Create, select, view, and edit schemas directly from the experiment configuration UI

Get Started