TTS-for-GPT-SoVITS: GPT-SoVITS Inference Plugin

Welcome to GSVI, an inference-specialized plugin built on top of GPT-SoVITS to enhance your text-to-speech (TTS) experience with a user-friendly API interface. This plugin enriches the original GPT-SoVITS project, making voice synthesis more accessible and versatile.

Features

High-level abstract interface for easy character and emotion selection
Comprehensive TTS engine support (speaker selection, speed adjustment, volume control)
User-friendly design for everyone
High compatibility and extensibility for various platforms and applications (for example: SillyTavern)

Installation

Self-Deployment

Use our optimized fork, GSVI on GitHub, for extended functionalities and plugin compatibility. Follow the installation instructions provided.

Prezip (For Windows)

Windows users can use our prezip, which includes pre-trained models, a Python environment, and a launcher written in Easy-Programming-Language. Download the prezip and follow the installation guide on our Yuque documentation page.

Usage

TTS

Start with a single gradio file

Gradio Application: src/Exhibition_Webui.py

Start with backend and frontend mod

Flask Backend Program: src/tts_backend.py
Gradio Frontend Application: src/TTS_Webui.py
Other Frontend Applications or Services Using Our API

Manage Tools

Gradio Model Management Interface: src/Character_Manager.py

For API documentation, visit our Yuque documentation page.

Getting Started

Read our documentation and usage instructions before starting.
Go and see our huggingface demo
If you encounter issues, join our community or consult the FAQ. QQ Group: 863760614

We look forward to seeing how you use GSVI to bring your creative projects to life!