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

# Music & Sound Effects

> Generate music and sound effects with Venice's asynchronous audio API: choose a model, quote the cost, queue a job, and download the completed audio.

Music and sound-effect generation is asynchronous. Choose a model, request a price quote, queue the generation, then poll until Venice returns the finished audio file.

## Choose a model

Browse [Music & Sound Effects Models](/models/music) for current model IDs, pricing, duration limits, and supported features.

You can also discover model capabilities at runtime:

```bash theme={"system"}
curl "https://api.venice.ai/api/v1/models?type=music" \
  -H "Authorization: Bearer $VENICE_API_KEY"
```

Check each model's metadata before setting optional fields such as `duration_seconds`, `lyrics_prompt`, `force_instrumental`, or `loop`. Unsupported fields cause an HTTP `400` response.

## Generation flow

| Endpoint                                                         | Purpose                                   |
| ---------------------------------------------------------------- | ----------------------------------------- |
| [`POST /audio/quote`](/api-reference/endpoint/audio/quote)       | Estimate the generation cost in USD       |
| [`POST /audio/queue`](/api-reference/endpoint/audio/queue)       | Start a music or sound-effect generation  |
| [`POST /audio/retrieve`](/api-reference/endpoint/audio/retrieve) | Poll the job and download completed audio |
| [`POST /audio/complete`](/api-reference/endpoint/audio/complete) | Delete stored media after downloading it  |

## 1. Get a price quote

Quote the request before generating media. Include the same model and duration you plan to send to the queue endpoint.

```bash theme={"system"}
curl https://api.venice.ai/api/v1/audio/quote \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs-music",
    "duration_seconds": 30
  }'
```

The response contains the estimated cost in USD:

```json theme={"system"}
{
  "quote": 0.75
}
```

## 2. Queue the generation

<CodeGroup>
  ```bash Music theme={"system"}
  curl https://api.venice.ai/api/v1/audio/queue \
    -H "Authorization: Bearer $VENICE_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "elevenlabs-music",
      "prompt": "Warm cinematic strings with a gentle piano melody, hopeful and spacious",
      "duration_seconds": 30,
      "force_instrumental": true
    }'
  ```

  ```bash Sound effect theme={"system"}
  curl https://api.venice.ai/api/v1/audio/queue \
    -H "Authorization: Bearer $VENICE_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "elevenlabs-sound-effects-v2",
      "prompt": "Ocean waves rolling onto a pebble beach at night",
      "duration_seconds": 10
    }'
  ```
</CodeGroup>

A successful request returns the model and a queue ID:

```json theme={"system"}
{
  "model": "elevenlabs-music",
  "queue_id": "123e4567-e89b-12d3-a456-426614174000",
  "status": "QUEUED"
}
```

Save both `model` and `queue_id`; the retrieve and complete endpoints require them.

## 3. Poll and download

Call `/audio/retrieve` with the values from the queue response:

```bash theme={"system"}
curl https://api.venice.ai/api/v1/audio/retrieve \
  -H "Authorization: Bearer $VENICE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs-music",
    "queue_id": "123e4567-e89b-12d3-a456-426614174000"
  }' \
  --output response.bin
```

Inspect the response `Content-Type`:

| Content-Type                               | Meaning                        | Action                                           |
| ------------------------------------------ | ------------------------------ | ------------------------------------------------ |
| `application/json`                         | Generation is still processing | Read the timing fields, wait, and poll again     |
| `audio/mpeg`, `audio/wav`, or `audio/flac` | Generation is complete         | Save the binary body with the matching extension |

A processing response looks like this:

```json theme={"system"}
{
  "status": "PROCESSING",
  "average_execution_time": 20000,
  "execution_duration": 5200
}
```

Both timing values are milliseconds.

## Complete example

This Python example queues instrumental music, polls every five seconds, and saves the result with an extension based on its content type.

```python theme={"system"}
import os
import time
from pathlib import Path

import requests

BASE_URL = "https://api.venice.ai/api/v1"
HEADERS = {
    "Authorization": f"Bearer {os.environ['VENICE_API_KEY']}",
    "Content-Type": "application/json",
}

generation = {
    "model": "elevenlabs-music",
    "prompt": "Warm cinematic strings with a gentle piano melody, hopeful and spacious",
    "duration_seconds": 30,
    "force_instrumental": True,
}

quote = requests.post(f"{BASE_URL}/audio/quote", json={
    "model": generation["model"],
    "duration_seconds": generation["duration_seconds"],
})
quote.raise_for_status()
print(f"Estimated cost: ${quote.json()['quote']:.2f}")

queued = requests.post(f"{BASE_URL}/audio/queue", headers=HEADERS, json=generation)
queued.raise_for_status()
job = queued.json()

content_type_to_extension = {
    "audio/mpeg": ".mp3",
    "audio/wav": ".wav",
    "audio/flac": ".flac",
}

while True:
    result = requests.post(
        f"{BASE_URL}/audio/retrieve",
        headers=HEADERS,
        json={"model": job["model"], "queue_id": job["queue_id"]},
    )
    result.raise_for_status()
    content_type = result.headers.get("Content-Type", "").split(";")[0]

    if content_type in content_type_to_extension:
        output = Path("generated-audio" + content_type_to_extension[content_type])
        output.write_bytes(result.content)
        print(f"Saved {output}")
        break

    status = result.json()
    print(f"Status: {status['status']}")
    time.sleep(5)

requests.post(
    f"{BASE_URL}/audio/complete",
    headers=HEADERS,
    json={"model": job["model"], "queue_id": job["queue_id"]},
).raise_for_status()
```

<Note>
  The quote endpoint does not require authentication, but queue, retrieve, and complete requests do.
</Note>

## Prompting tips

* For music, describe genre, instruments, mood, tempo, structure, and whether vocals are desired.
* For sound effects, describe the source, environment, intensity, timing, and perspective.
* Use `lyrics_prompt` only when the selected model supports lyrics.
* Use `force_instrumental` or `loop` only when the model metadata reports support.

## Related resources

* [Music & Sound Effects Models](/models/music)
* [Queue Audio Generation API](/api-reference/endpoint/audio/queue)
* [Retrieve Audio API](/api-reference/endpoint/audio/retrieve)