> ## Documentation Index
> Fetch the complete documentation index at: https://deepl-c950b784-docs-pipeline-20260820-095624.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Translate an Audio File with the Voice Translate Job API

> Submit a pre-recorded audio file for translation and download the results using the asynchronous Voice Translate Job API.

The Voice Translate Job API translates pre-recorded audio files asynchronously. You submit a job, upload your audio file, poll for results, and download each output when it's ready. This guide walks through all four steps with a complete curl example.

Use this API for batch or offline workloads: podcasts, meeting recordings, video files, and similar pre-recorded content. For live audio, use the [real-time Voice API](/docs/voice/overview).

<Warning>
  This API is only available to select DeepL customers and may change without notice. See [alpha and beta features](/docs/resources/alpha-and-beta-features) for details, or contact your customer success manager to request access.
</Warning>

## Prerequisites

* A DeepL API account with Voice Translate Job API access
* Your DeepL API key (set as `DEEPL_API_KEY` in the examples below)
* A pre-recorded audio file in a [supported format](/api-reference/jobs-voice-translate/reference#supported-source-audio-formats)

For file size, duration, concurrency, and supported format constraints, see the [Translate Audio Files reference](/api-reference/jobs-voice-translate/reference#limits).

API Pro users call `https://api.deepl.com`. API Free users call `https://api-free.deepl.com` instead.

## The four-step workflow

Every translation follows the same pattern:

1. **Create a job** to declare your source file and desired outputs. The API returns an upload URL.
2. **Upload your audio file** directly to that URL.
3. **Poll for status** until all targets are complete (or failed).
4. **Download each result** from its download URL.

The sections below walk through each step.

## Step 1: Create a job

Send a POST request to `/v1/jobs/voice/translate` with the source file metadata and your desired translation targets.

The `content_length` and `content_type` fields in `source_file` must match your actual file exactly. The API uses them to provision a pre-signed upload URL, so mismatches will cause the upload in step 2 to fail.

```bash theme={null}
curl -X POST "https://api.deepl.com/v1/jobs/voice/translate" \
  -H "Authorization: DeepL-Auth-Key $DEEPL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "source_file": {
      "name": "podcast-episode-42.mp3",
      "content_type": "audio/mpeg",
      "content_length": 15728640
    },
    "parameters": {
      "source_language": "en"
    },
    "targets": [
      { "language": "de", "type": "text/plain" },
      { "language": "es", "type": "audio/pcm;encoding=s16le;rate=16000" }
    ]
  }'
```

A successful response includes a `job_id`, an `upload_url`, and a `signature`:

```json theme={null}
{
  "job_id": "a74d88fb-ed2a-4943-a664-a4512398b994",
  "signature": "eyJhbGciOiJIUzI1NiIs...",
  "upload_url": "https://assets.deepl.com/collections/a74d88fb-ed2a-4943-a664-a4512398b994/assets/b1c2d3e4-f5a6-7890-abcd-ef1234567890"
}
```

Save the `job_id` for polling and the `upload_url` for the next step. You have 5 minutes to complete the upload before the URL expires.

Each job can have multiple targets in different output types, so a single English recording can produce a German transcript, French subtitles, and Spanish audio in one request. See the [Translate Audio Files reference](/api-reference/jobs-voice-translate/reference) for the full list of output types and supported languages.

## Step 2: Upload your audio file

PUT your audio file to the `upload_url` returned in step 1. Set `Content-Type` to the same value you declared in `source_file.content_type`.

Do not include your DeepL API key in the upload request. The URL is pre-signed and only requires the `Content-Type` header. Adding an `Authorization` header will cause the request to fail.

```bash theme={null}
curl -X PUT "https://assets.deepl.com/collections/a74d88fb-ed2a-4943-a664-a4512398b994/assets/b1c2d3e4-f5a6-7890-abcd-ef1234567890" \
  -H "Content-Type: audio/mpeg" \
  --data-binary @podcast-episode-42.mp3
```

A `200 OK` response with no body means the upload succeeded. Processing starts automatically once the file is received.

## Step 3: Poll for status

Check the job status by sending a GET request to `/v1/jobs/voice/translate/{job_id}`. Each target in the `results` array has its own `status` field that progresses independently.

```bash theme={null}
curl "https://api.deepl.com/v1/jobs/voice/translate/a74d88fb-ed2a-4943-a664-a4512398b994" \
  -H "Authorization: DeepL-Auth-Key $DEEPL_API_KEY"
```

While targets are still processing, the response looks like this:

```json theme={null}
{
  "job_id": "a74d88fb-ed2a-4943-a664-a4512398b994",
  "operation": "translate",
  "product": "voice",
  "parameters": { "source_language": "en" },
  "source_file": {
    "name": "podcast-episode-42.mp3",
    "content_type": "audio/mpeg",
    "content_length": 15728640
  },
  "targets": [
    { "language": "de", "type": "text/plain" },
    { "language": "es", "type": "audio/pcm;encoding=s16le;rate=16000" }
  ],
  "results": [
    { "status": "processing" },
    { "status": "processing" }
  ],
  "created_at": "2026-10-01T01:03:03.444Z",
  "updated_at": "2026-10-01T04:03:03.333Z"
}
```

When a target completes, its result entry includes a `download_url` and `signature`:

```json theme={null}
{
  "results": [
    {
      "status": "complete",
      "download_url": "https://assets.deepl.com/collections/a74d88fb/assets/c3d4e5f6",
      "signature": "eyJhbGciOiJIUzI1NiIs..."
    },
    {
      "status": "failed",
      "error": { "message": "processing failed" }
    }
  ]
}
```

Results are returned in the same order as the `targets` array in your create request, so `results[0]` corresponds to `targets[0]`.

Poll every 10-30 seconds for files under 100 MB. Larger files or longer recordings may take several minutes. Keep polling until every result has reached a terminal status: `complete`, `failed`, or `downloaded`.

## Step 4: Download results

For each target with `status: complete`, download the result from its `download_url`. No authentication header is needed — the URL is pre-signed.

```bash theme={null}
# Download the German plain-text transcript
curl -o translation_de.txt \
  "https://assets.deepl.com/collections/a74d88fb/assets/c3d4e5f6"
```

After you download a result, its status transitions to `downloaded` and the asset is queued for deletion. You have 1 hour from the time the upload completes to download all results. After that window, or once all results are downloaded, the job is deleted and returns `404`.

## Handling partial failures

Individual targets can fail while others succeed. Check each result's `status` independently and handle `failed` results by reading the `error.message` field. You cannot retry a failed target within an existing job; you need to create a new job.