Create Translation

Brief: Translate audio into target text using OpenAI Whisper family models.

---

Overview

• Method: POST
• Path: /v1/audio/translations
• Content-Type: multipart/form-data

Authentication

• Header: Authorization: Bearer <token>
• Supports bearer token authentication

---

Request Body Parameters

Parameter	Type	Required	Description
file	file	Yes	Audio file object, supporting flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm
model	string	Yes	Model name, typically whisper-1
prompt	string	No	Optional prompt text to guide translation style. Should be written in English
response_format	string	No	Output format. Supported values: json, text, srt, verbose_json, vtt. Default: json
temperature	number	No	Sampling temperature, range 0 to 1. Higher values produce more random output, lower values produce more stable output
timestamp_granularities	array	No	Timestamp granularity. Options: response_format, verbose_json, word, segment
language	string	No	Audio language. Use ISO-639-1 codes to improve translation accuracy

---

curl Example

curl -X POST "https://api.gpt.ge/v1/audio/translations" \
  -H "Authorization: Bearer sk-xxxx" \
  -F "file=@./audio.wav" \
  -F "model=whisper-1" \
  -F "response_format=json" \
  -F "language=zh"

JavaScript (fetch) Example

const formData = new FormData();
formData.append('file', audioFile);
formData.append('model', 'whisper-1');
formData.append('response_format', 'json');
formData.append('language', 'zh');

fetch('https://api.gpt.ge/v1/audio/translations', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer sk-xxxx'
  },
  body: formData
}).then(r => r.json()).then(console.log);

Python (requests) Example

import requests

with open('audio.wav', 'rb') as f:
    files = {'file': f}
    data = {
        'model': 'whisper-1',
        'response_format': 'json',
        'language': 'zh'
    }
    response = requests.post(
        'https://api.gpt.ge/v1/audio/translations',
        headers={'Authorization': 'Bearer sk-xxxx'},
        files=files,
        data=data
    )
print(response.json())

---

Response Example (200)

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

Note: The translation endpoint returns translated text. Selecting response_format as srt, vtt, or verbose_json will also include timestamp or subtitle information.