MiniMax Voice Cloning
Model Information
Display Name: MiniMax Voice Cloning
API Model ID: minimax/voice-cloning
Category: Voice Management
Description: MiniMax Voice Cloning creates a reusable voice profile from a short audio sample. Once cloned, the voice can be used with MiniMax Speech 2.8 HD to generate speech in that voice with full emotion and style control. ## Key Features - **Instant cloning** — only 10 seconds to 5 minutes of reference audio needed - **High fidelity** — preserves voice timbre, accent, and natural characteristics - **Noise reduction** — optional built-in noise reduction for noisy samples - **Volume normalization** — optional automatic level adjustment - **Reusable voice_id** — clone once, use unlimited times with Speech 2.8 HD ## How It Works 1. Send an audio file (MP3, M4A, or WAV, 10s–5min, max 20MB) 2. Receive a `voice_id` and a preview audio sample 3. Use the `voice_id` with `minimax/speech-28-hd` to generate speech in the cloned voice ## Parameters | Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | audio | URL | Yes | — | Audio file to clone (MP3, M4A, WAV). 10s to 5min, max 20MB | | accuracy | number | No | 0.7 | Validation accuracy threshold (0–1) | | model | string | No | speech-02-hd | Target model: speech-02-hd or speech-02-turbo | | need_noise_reduction | boolean | No | false | Enable noise reduction for noisy samples | | need_volume_normalization | boolean | No | false | Enable automatic volume leveling | ## Output Returns JSON with `voice_id` (for use in TTS), `preview` (audio URL), and `model` (trained model name).
How to Use This Model
To use MiniMax Voice Cloning via the HInow.ai API, use the model ID: minimax/voice-cloning
API Request Example
POST https://api.hinow.ai/v1/chat/completions
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
{
"model": "minimax/voice-cloning",
"messages": [
{"role": "user", "content": "Your input here"}
]
}
Pricing
- total: $1.00
Available Parameters
- audio: Audio file URL to clone (MP3, M4A, WAV). 10s to 5min, max 20MB
- accuracy: Validation accuracy threshold (0-1). Default: 0.7
- model: Target model: speech-02-hd or speech-02-turbo
- need_noise_reduction: Enable noise reduction for noisy samples
- need_volume_normalization: Enable volume normalization
Quick Reference
To use this model, set: "model": "minimax/voice-cloning"
Featured: No
Documentation: https://hinow.ai/models/minimax/voice-cloning
API Endpoint: https://api.hinow.ai/v1
