Create Voice Cloning Task

Start a voice cloning task. You can upload audio files directly or provide URLs to existing audio files.

Request Information

Method: POST
Endpoint: /api/instant/create-task
Content Type: multipart/form-data or application/json

Request Parameters

Parameter	Type	Required	Description
`audio`	File (binary)	Yes*	Audio file for voice cloning. Supported formats include WAV, MP3, and M4A. Only supports multipart/form-data format. You must provide either `audio` or `audio_url`.
`audio_url`	string	Yes*	Publicly accessible audio file URL (WAV, MP3, M4A). Supports both multipart/form-data and application/json formats. You must provide either `audio` or `audio_url`.
`text`	string	Yes	The text you want to synthesize with the cloned voice.
`api_key`	string	Yes	Your unique API key for authentication and access. This key is used to verify your request and link it to your user account.
`model`	string	No	Voice cloning model version. Options: `v1`, `v2`, `v-mul`. Default is `v2`. Different models vary in audio quality, processing speed, and multilingual support.
`speed_ratio`	float	No	Speed ratio, range 0.5-2.0, default is 1.0
`pitch_ratio`	float	No	Pitch offset, range -10 to 10 semitones, default is 0
`volume_ratio`	float	No	Volume ratio, range 0.1-2.0, default is 1.0
`emotion_control`	object	No	Emotion control parameters (V2 model only). Supports multiple emotion control modes, see below for details

*Note:

Either audio or audio_url parameter is required, at least one must be provided
When using application/json format, only audio_url parameter is supported
When using multipart/form-data format, both parameters are supported

Response

Success Response


{
  "task_id": "1406bf34-735c-4b21-98ac-a135b2afb1c8",
  "status": "pending"
}

Error Response

400 Bad Request: Missing required parameters (e.g., api_key, or neither audio nor audio_url provided)

Example Requests

Using Audio File


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -F "[email protected]" \
  -F "text=This is a long text suitable for async interface processing..." \
  -F "api_key=your_api_key_here"

Using Audio URL


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -F "audio_url=https://example.com/sample.mp3" \
  -F "text=This is a long text suitable for async interface processing..." \
  -F "api_key=your_api_key_here"

Using Specified Model


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -F "[email protected]" \
  -F "text=This is a long text suitable for async interface processing..." \
  -F "api_key=your_api_key_here" \
  -F "model=v2"

Using Audio Processing Parameters


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -F "[email protected]" \
  -F "text=This is a long text suitable for async interface processing..." \
  -F "api_key=your_api_key_here" \
  -F "model=v2" \
  -F "speed_ratio=1.2" \
  -F "pitch_ratio=2" \
  -F "volume_ratio=1.5"

Using JSON Format (with audio_url)


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "https://example.com/sample.mp3",
    "text": "This is a long text suitable for async interface processing...",
    "api_key": "your_api_key_here",
    "model": "v2",
    "speed_ratio": 1.2,
    "pitch_ratio": 2,
    "volume_ratio": 1.5
  }'

Model Selection Guide

The API supports three voice cloning models that you can choose from based on your needs:

Model	Description	Use Cases
`v2`	Second generation model (default), balances audio quality and processing speed	Recommended for most scenarios, provides high-quality voice cloning results
`v1`	First generation model, faster processing speed	Suitable for scenarios with high processing speed requirements
`v-mul`	Multilingual model, supports cross-language voice cloning	Suitable for applications requiring multilingual support

Note: If the model parameter is not specified, the system will use the v2 model by default.

Emotion Control Parameters (emotion_control)

V2 model only supports emotion control functionality through the emotion_control object. This can be passed directly in JSON format requests.

Emotion Control Modes

Mode	mode Value	Description	Additional Parameters
Same as Reference	`same_as_reference`	Use the emotion from reference audio (default)	None
Reference Audio	`reference_audio`	Use specified reference audio to control emotion	`reference_audio_url`
Vector Control	`vector`	Use 8-dimensional emotion vector to control emotion	`vector` object
Text Description	`text`	Use text description to control emotion	`text` string
Random Emotion	`random`	Randomly generate emotion	None

8-Dimensional Emotion Vector (vector mode)

When using vector mode, you can precisely control the intensity of 8 emotions, each value ranges from 0.0 to 1.0:


{
  "joy": 0.5,        // Joy (0.0-1.0)
  "anger": 0.0,      // Anger (0.0-1.0)
  "sorrow": 0.0,     // Sorrow (0.0-1.0)
  "fear": 0.0,       // Fear (0.0-1.0)
  "excitement": 0.3, // Excitement (0.0-1.0)
  "depression": 0.0, // Depression (0.0-1.0)
  "surprise": 0.2,   // Surprise (0.0-1.0)
  "calm": 0.0        // Calm (0.0-1.0)
}

Emotion Control Examples

Using Vector Control (Happy + Excited)


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "https://example.com/sample.mp3",
    "text": "What a beautiful day!",
    "api_key": "your_api_key_here",
    "model": "v2",
    "emotion_control": {
      "mode": "vector",
      "vector": {
        "joy": 0.8,
        "excitement": 0.6,
        "surprise": 0.3,
        "calm": 0.2
      }
    }
  }'

Using Text Description Control


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "https://example.com/sample.mp3",
    "text": "This is absolutely amazing!",
    "api_key": "your_api_key_here",
    "model": "v2",
    "emotion_control": {
      "mode": "text",
      "text": "excited and happy"
    }
  }'

Using Reference Audio Control


curl -X POST https://aivoiceclonefree.com/api/instant/create-task \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "https://example.com/sample.mp3",
    "text": "I feel very happy",
    "api_key": "your_api_key_here",
    "model": "v2",
    "emotion_control": {
      "mode": "reference_audio",
      "reference_audio_url": "https://example.com/emotion-ref.wav"
    }
  }'

Task Status Description

After creating a task, you will receive a task_id and initial status pending. Possible status values:

Status	Description
`pending`	Task submitted, waiting for processing
`processing`	Task is being processed
`completed`	Task completed
`failed`	Task processing failed

Next Steps

After successful task creation:

Save the returned task_id
Use Task Status Query to monitor progress
Download audio via Get Task Result after completion

Best Practices

Text Length: Recommended single task text length should not exceed 10,000 characters
Audio Quality: Use high-quality audio samples for better cloning results
Request Format:
- Using multipart/form-data: Can directly upload audio file (audio parameter) or use audio URL (audio_url parameter)
- Using application/json: Can only use audio URL (audio_url parameter)
API Limits: Be aware of API call frequency limits, avoid overly frequent requests