Create Translation

Translates audio into English text

Overview

Translates audio in any supported language into English text. Unlike transcription, this endpoint always outputs English text regardless of the input language.

This page documents audio translation (POST /v1/audio/translations). For text translation, use POST /v1/translations.

Do not use recommended_for=translation for this endpoint. That recommendation scene is reserved for text translation models on POST /v1/translations.

Request Body

file · file · required

The audio file to translate. Supported formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, webm. Maximum file size is 25 MB.

model · string · default: "whisper-1"

The model to use. Currently only whisper-1 is supported.

prompt · string

An optional text to guide the model’s style or continue a previous segment. Should be in English.

response_format · string · default: "json"

The format of the output. Options: json, text, srt, verbose_json, vtt.

temperature · number

The sampling temperature, between 0 and 1. Higher values like 0.8 produce more random output, while lower values like 0.2 make output more focused and deterministic.

Response

text · string

The translated text in English.

For verbose_json format, the response also includes:

language · string

The detected language of the input audio.

duration · number

The duration of the input audio in seconds.

segments · array

Segments of the translated text with timestamps.

cURL

curl -X POST "https://api.lemondata.cc/v1/audio/translations" \
  -H "Authorization: Bearer sk-your-api-key" \
  -F "file=@german_audio.mp3" \
  -F "model=whisper-1"

json

{
  "text": "Hello, my name is Wolfgang and I come from Germany. Where are you from?"
}

Translation vs Transcription

Feature	Translation	Transcription
Output language	Always English	Same as input
Use case	Convert foreign audio to English	Preserve original language
Language parameter	Not applicable	Optional hint

The translation endpoint automatically detects the source language and translates to English. The language parameter from transcription is ignored.