Evaluate and Select Best AI Output from Multiple Options

Eval Select Best is a promptfoo example workflow that demonstrates evaluation capabilities within the promptfoo framework. It provides a starting point for running evaluations using promptfoo's testing infrastructure.

Use this workflow when you want to learn how promptfoo's evaluation features work through a concrete example. It's a reference implementation that shows the basic structure and commands needed to run promptfoo evaluations.

Do not use this if you need a production-ready evaluation setup - this is an example workflow meant for learning and experimentation, not a complete solution.

You provide a promptfooconfig.yaml configuration file and set your OPENAI_API_KEY environment variable before running.

You receive evaluation results that can be viewed through the promptfoo interface after running the evaluation commands.

This example integrates with OpenAI's API (requires OPENAI_API_KEY environment variable) and uses the promptfoo evaluation framework for running tests and viewing results.

This workflow is built for prompt engineers, AI application developers, and ML teams who want to explore promptfoo's evaluation capabilities through a working example.

Getting started

Install and run the example with these commands:

npx promptfoo@latest init --example eval-select-best
cd eval-select-best

Then execute the evaluation:

promptfoo eval

After evaluation completes, view results by running promptfoo view.