​
Introduction to NANSY Model

NANSY(Neural Analysis and Synthesis) model-based voice AI technology from Supertone generates high-quality voices that are difficult to distinguish from real human voices. NANSY is an integrated neural network framework designed to perform voice-related generation tasks. This model serves as the foundation for various downstream tasks such as voice and song synthesis, voice conversion, and voice design. Through its integrated structure, it maintains consistent voice characteristics during the generation process, and can express anyone’s voice through the control of four individual elements.

​
How to Generate Voice

To generate Supertone’s high-quality AI voices, you need to use the Supertone API. The voice generation process through the API is as follows:

1

Get API Key

After signing up for Supertone API service, please apply for closed beta at the Console Page. Once your application is approved, you can get your API Key from the console page.

2

Select Voice

You can either call the Get Voices API to view the list of available voices or sign up for Supertone Play to test all voices free for 2 weeks. Once you find a voice you like, copy its ID and input it as an API call parameter.

3

Generate Voice

You can generate AI voice from text by calling Supertone’s Text-to-speech API.

4

Use the Results

Download the generated voice file or play it via streaming. You can use it for content creation and various other applications.

To start using Supertone API right away, please check the Quick Start page.

​
Key Features

​
1. High-Quality Voice Synthesis

Supertone’s AI voice synthesis technology provides natural intonation and rich emotional expression.

Key Features

  1. Generate voices with natural intonation and prosody
  2. Express various emotions and nuances
  3. Choose file formats according to user preferences
    • wav: For lossless high-quality audio needs
    • mp3: For efficient file size in general use

​
2. Multi-language Support

Supertone API provides support for various languages for global services.

Supported Languages

  1. Korean(ko)
  2. Japanese(ja)
  3. English(en)

Each language provides pronunciation and intonation optimized for that language, and we plan to continuously expand our language support.

​
3. Various Voices

We provide voices suitable for various characters and situations through our rich voice portfolio.

Voice Classification Tags

  1. Gender
    • Male
    • Female
  2. Age Group
    • Child
    • Young
    • Middle-aged
    • Elderly
  3. Style
    • Support for voice-specific emotion expression
    • Various speaking styles suitable for different situations and content
    • Unique timbres that embody character traits

Each voice can express unique characteristics and emotions matching its character, and you can test the actual voices free for 2 weeks by signing up for Supertone Play service.

WelcomeText-to-speech Guide