Audio · OpenAI

Audio Translation

Translate audio in any language to English text using OpenAI's Whisper model.

POST/v1/audio/translations

Supported Models

ModelProvider
whisper-1OpenAI
The translation endpoint always translates to English. For speech-to-text in the original language, use the Transcription endpoint instead.

Request

Multipart Form Parameters

filefileformrequired

Audio file (mp3, mp4, mpeg, mpga, m4a, wav, webm — max 25 MB)

modelstringformrequired

Model ID — use "whisper-1"

promptstringform

Hint text to guide translation style

response_formatstringform

Output format

Default: json

Options: json, text, srt, verbose_json, vtt

temperaturenumberform

Sampling temperature (0-1)

Default: 0

cURL
curl https://api.metriqual.com/v1/audio/translations \
  -H "Authorization: Bearer mql_your_key" \
  -F file=@german_audio.mp3 \
  -F model=whisper-1
TypeScript SDK
const result = await mql.audio.translate({
  file: audioFile,
  model: 'whisper-1',
});

console.log(result.text); // English translation
Python SDK
with open("german_audio.mp3", "rb") as f:
    result = mql.audio.translate(file=f, model="whisper-1")
print(result["text"])  # English translation

Response

Returns the translated English text. Use verbose_json for timestamps and segments.

200
json (default)
{
  "text": "Hello, how are you? I am doing well, thank you."
}
200
verbose_json
{
  "text": "Hello, how are you?",
  "language": "german",
  "duration": 4.21,
  "segments": [
    {
      "id": 0,
      "start": 0.0,
      "end": 4.21,
      "text": " Hello, how are you?",
      "temperature": 0
    }
  ]
}
Next
Quick Start