promptfoo/examples/openai-audio

OpenAI Audio Example

This example demonstrates how to use promptfoo to test OpenAI's audio capabilities using audio-capable models. The example focuses on audio input processing (speech-to-text) and audio output generation (speech-to-speech).

Quick Start

You can initialize this example in a new directory using:

npx promptfoo@latest init --example openai-audio

This will create all necessary files and folder structure to get started quickly.

Setup

1. Set your OpenAI API key as an environment variable:

export OPENAI_API_KEY=your-api-key-here

2. The example includes sample audio files in the assets directory:
   • Armstrong_Small_Step.mp3 - Neil Armstrong's moon landing speech
   • Kennedy_berliner.mp3 - JFK's "Ich bin ein Berliner" speech

Files

• promptfooconfig.yaml: Configuration file defining the providers and tests
• audio-input.json: JSON template for the audio input prompt

Running the Example

From the root directory of promptfoo, run:

npx promptfoo eval -c examples/openai-audio/promptfooconfig.yaml