Prompt Playground

The Prompt Playground is where you iterate on prompts and test them quickly. It supports various AI providers (OpenAI, Anthropic, Gemini, Azure) and custom model endpoints. Use it as a prompt IDE to build, experiment with, and evaluate prompts and models for your task.

• Speed: Quickly test variations in the prompt, model, invocation parameters, tools, and output format.

• Reproducibility: Phoenix records all playground runs as traces and experiments. You can then annotate and evaluate them.

• Datasets: Use dataset examples as fixtures to test a prompt variant thoroughly and evaluate it systematically.

• Prompt Management: Load, edit, and save prompts in the playground.

To use the playground, see Using the Playground.