> ## Documentation Index
> Fetch the complete documentation index at: https://docs.voidai.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Create Translation

> Submit audio to the VoidAI translation endpoint with multipart upload limits and supported parameters.

`POST https://api.voidai.app/v1/audio/translations`

Requires [authentication](/authentication). Send `multipart/form-data` with a non-empty `file`, at most **25 MiB** (26,214,400 bytes). Choose a model that advertises this endpoint; the default model is `whisper-1`.

| Field | Type | Description |
| - | - | - |
| `file` | file, required | Audio to process. Accepted encodings depend on the selected model/provider. |
| `model` | string | Exact model ID; defaults to `whisper-1`. |
| `prompt` | string | Optional context. |
| `response_format` | string | `json` (default), `text`, `srt`, `verbose_json`, or `vtt`. |
| `temperature` | number | The gateway accepts values from 0 to 2; the selected provider may impose narrower limits. |

```bash theme={null}
curl https://api.voidai.app/v1/audio/translations \
  -H "Authorization: Bearer $VOIDAI_API_KEY" \
  -F model=whisper-1 \
  -F file=@audio.mp3 \
  -F response_format=json
```

For the JSON format, read the transcript from `text`. Output format and translation behavior depend on the selected model/provider. Consult [models and current prices](https://voidai.app/models).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.