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

# Audio APIs

> Convert text to natural-sounding speech with Melo TTS

# Audio APIs

<Warning>
  **Sunset Notice:** Melo TTS is being discontinued. This model will be removed in a future update. Please plan your migration accordingly.
</Warning>

Convert text to natural-sounding speech using Melo TTS. Generate high-quality audio in multiple languages with customizable speed and speaker options.

## Text-to-Speech

### Endpoint

```
POST https://api.hyperbolic.xyz/v1/audio/generation
```

### Basic Example

<Tabs>
  <Tab title="Python">
    ```python theme={null}
    import requests
    import base64

    url = "https://api.hyperbolic.xyz/v1/audio/generation"
    headers = {
        "Content-Type": "application/json",
        "Authorization": "Bearer YOUR_API_KEY"
    }
    data = {
        "text": "Welcome to Hyperbolic! This is an example of text-to-speech generation.",
        "speed": 1
    }

    response = requests.post(url, headers=headers, json=data)
    result = response.json()

    # Decode and save the audio file
    audio_data = base64.b64decode(result["audio"])
    with open("output.mp3", "wb") as f:
        f.write(audio_data)
    ```
  </Tab>

  <Tab title="cURL">
    ```bash theme={null}
    curl -X POST "https://api.hyperbolic.xyz/v1/audio/generation" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer YOUR_API_KEY" \
      -d '{
        "text": "Welcome to Hyperbolic! This is an example of text-to-speech generation.",
        "speed": 1
      }' | jq -r ".audio" | base64 -d > output.mp3
    ```
  </Tab>
</Tabs>

## Request Parameters

### Required Parameters

| Parameter | Type   | Description                   |
| --------- | ------ | ----------------------------- |
| `text`    | string | The text to convert to speech |

### Optional Parameters

| Parameter       | Type   | Default | Description                                                               |
| --------------- | ------ | ------- | ------------------------------------------------------------------------- |
| `language`      | string | `EN`    | Language code for speech generation                                       |
| `speaker`       | string | -       | Speaker variant (language-specific)                                       |
| `speed`         | float  | 1       | Speech speed multiplier (0.1-5)                                           |
| `sdp_ratio`     | float  | -       | Stochastic duration predictor ratio, controls prosody variation (0-1)     |
| `noise_scale`   | float  | -       | Controls randomness in speech generation for more natural variation (0-1) |
| `noise_scale_w` | float  | -       | Controls randomness in duration prediction for timing variation (0-1)     |

## Response Format

The API returns a JSON object containing base64-encoded MP3 audio:

```json theme={null}
{
  "audio": "SUQzBAAAAAAAI1RTU0UAAAAPAAADTGF2ZjU4Ljc2..."
}
```

### Decoding the Response

```python theme={null}
import base64

def save_audio(response_json, filename="output.mp3"):
    """Decode and save the generated audio."""
    audio_data = base64.b64decode(response_json["audio"])
    with open(filename, "wb") as f:
        f.write(audio_data)
```

## Supported Languages

Melo TTS supports 6 languages with various speaker options:

| Language | Code | Available Speakers                    |
| -------- | ---- | ------------------------------------- |
| English  | `EN` | `EN-US`, `EN-BR`, `EN-INDIA`, `EN-AU` |
| Spanish  | `ES` | `ES`                                  |
| French   | `FR` | `FR`                                  |
| Chinese  | `ZH` | `ZH`                                  |
| Japanese | `JP` | `JP`                                  |
| Korean   | `KR` | `KR`                                  |

<Info>
  English has multiple speaker variants for different accents: US (American), BR (British), INDIA (Indian), and AU (Australian).
</Info>

## Using Different Languages and Speakers

<Tabs>
  <Tab title="Python">
    ```python theme={null}
    import requests
    import base64

    url = "https://api.hyperbolic.xyz/v1/audio/generation"
    headers = {
        "Content-Type": "application/json",
        "Authorization": "Bearer YOUR_API_KEY"
    }

    # Generate speech with Australian English accent
    data = {
        "text": "G'day! Welcome to our application.",
        "language": "EN",
        "speaker": "EN-AU",
        "speed": 1
    }

    response = requests.post(url, headers=headers, json=data)
    result = response.json()

    audio_data = base64.b64decode(result["audio"])
    with open("australian_greeting.mp3", "wb") as f:
        f.write(audio_data)
    ```
  </Tab>

  <Tab title="cURL">
    ```bash theme={null}
    curl -X POST "https://api.hyperbolic.xyz/v1/audio/generation" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer YOUR_API_KEY" \
      -d '{
        "text": "G'\''day! Welcome to our application.",
        "language": "EN",
        "speaker": "EN-AU",
        "speed": 1
      }' | jq -r ".audio" | base64 -d > australian_greeting.mp3
    ```
  </Tab>
</Tabs>

### Multilingual Example

```python theme={null}
# French
data = {
    "text": "Bonjour! Bienvenue sur notre plateforme.",
    "language": "FR",
    "speaker": "FR",
    "speed": 1
}

# Japanese
data = {
    "text": "こんにちは！ようこそ。",
    "language": "JP",
    "speaker": "JP",
    "speed": 1
}

# Chinese
data = {
    "text": "你好！欢迎使用我们的平台。",
    "language": "ZH",
    "speaker": "ZH",
    "speed": 1
}
```

## Speed Control

Adjust the `speed` parameter to control how fast the speech is generated:

| Speed  | Effect                      |
| ------ | --------------------------- |
| `0.5`  | Slow, deliberate speech     |
| `0.75` | Slightly slower than normal |
| `1.0`  | Normal speed (default)      |
| `1.5`  | Faster speech               |
| `2.0`  | Very fast                   |

### Example

```python theme={null}
# Slow narration for accessibility
data = {
    "text": "Please listen carefully to the following instructions.",
    "speed": 0.7
}

# Fast playback for quick review
data = {
    "text": "This is a quick summary of the key points.",
    "speed": 1.5
}
```

<Tip>
  Use slower speeds (0.5-0.8) for instructional content or accessibility needs. Use faster speeds (1.2-1.5) for content review or when listeners prefer quicker playback.
</Tip>

## Pricing

**Rate:** \$5.00 per 1 million characters

| Text Length          | Approximate Cost |
| -------------------- | ---------------- |
| 1,000 characters     | \$0.005          |
| 10,000 characters    | \$0.05           |
| 100,000 characters   | \$0.50           |
| 1,000,000 characters | \$5.00           |

<Info>
  There are no character limits per request. You are billed based on the total characters processed.
</Info>

## Use Cases

* **Voice assistants:** Add natural speech to chatbots and virtual assistants
* **Audiobook generation:** Convert written content to audio format
* **Accessibility:** Make content accessible for visually impaired users
* **Video narration:** Generate voiceovers for videos and presentations
* **Language learning:** Create pronunciation examples in multiple languages
* **Notification systems:** Generate audio alerts and announcements

## Next Steps

<CardGroup cols={3}>
  <Card title="Text APIs" icon="message" href="/docs/inference/text-apis">
    Generate text with large language models
  </Card>

  <Card title="Vision Language Models" icon="eye" href="/docs/inference/vlm-apis">
    Analyze images with multimodal AI
  </Card>

  <Card title="Image APIs" icon="image" href="/docs/inference/image-apis">
    Generate images from text prompts
  </Card>
</CardGroup>