AI Speech and Sound Generation Framework
1.5k 2026-04-18

OpenMOSS/MOSS-TTS

An open-source AI model family for high-fidelity, expressive speech and sound generation across diverse real-world applications.

Core Features

High-fidelity and high-expressiveness speech generation
Support for stable long-form speech and multi-speaker dialogue
Capabilities for voice/character design and environmental sound effects
Real-time streaming Text-to-Speech (TTS)
Efficient, PyTorch-free inference via llama.cpp and ONNX Runtime

Detailed Introduction

MOSS-TTS Family is an advanced open-source suite of AI models developed by MOSI.AI and the OpenMOSS team, specializing in high-fidelity and highly expressive speech and sound generation. It is meticulously designed to address complex real-world scenarios, offering robust solutions for stable long-form speech, multi-speaker dialogue, intricate voice and character design, realistic environmental sound effects, and efficient real-time streaming Text-to-Speech (TTS). The project emphasizes versatile and efficient deployment, notably supporting PyTorch-free inference via llama.cpp and ONNX Runtime, making it a powerful and adaptable tool for diverse applications ranging from content creation and gaming to interactive systems and accessibility.

OSS Alternative

Explore the best open source alternatives to commercial software.

© 2026 OSS Alternative. hotgithub.com - All rights reserved.