Principle:Openai Openai node Audio Translation
| Knowledge Sources | |
|---|---|
| Domains | Audio, Translation |
| Last Updated | 2026-02-15 00:00 GMT |
Overview
A principle for translating spoken audio in any supported language into English text using speech recognition and translation models.
Description
Audio Translation combines speech recognition with translation in a single step. It takes an audio file in any supported language and produces an English text transcription. Unlike transcription (which preserves the original language), translation always outputs English.
This is a convenience API that avoids the two-step process of transcribing in the original language and then translating the text.
Usage
Use this principle when you have non-English audio that needs to be converted to English text. For same-language transcription, use Audio Transcription instead.
Theoretical Basis
Audio translation follows a Audio → Model → English Text pipeline:
function translateAudio(audioFile, model, options):
response = await api.post('/audio/translations', multipart({
file: audioFile,
model: model,
prompt: options.prompt, // Context hint in English
response_format: options.format,
temperature: options.temperature,
}))
return response // English text translation