Multi-lingual large voice generation model, providing inference, training and deployment full-stack ability.

🍦 Speech-AI-Forge is a project developed around TTS generation model, implementing an API Server and a Gradio-based WebUI.

CosyVoice2 功能扩充（预训练音色推理/3s极速复刻/自然语言控制/自动识别/音色模型保存/API）

开箱即用的本地私有化部署语音服务，快速搭建Qwen3ASR/FunASR与CosyVoice2/3后端

CosyVoice_DPO_NOTES: Supercharge Your Cosyvoice model with Cutting-Edge DPO Fine-Tuning!

🎙️ CosyVoice All-in-One Docker - Production-ready TTS with Web UI, REST API & Voice Cloning

将 ZipEnhancer 模型从 ModelScope pipeline 中剥离，用纯 PyTorch 重新实现推理流程，并封装为高性能 FastAPI 批量语音降噪服务，为 CosyVoice 提供干净的人声输入

LightTTS is a lightweight TTS inference framework optimized for CosyVoice2 and CosyVoice3, enabling fast and scalable speech synthesis in Python and supports stream and bistream modes.

This repository provides a Docker image for CosyVoice

one component for text to sonic video in ComfyUI

Home Assistant integrates Alibaba Cloud's BaiLian Platform TTS

CosyVoice3 LoRA fine-tuning companion repo — PEFT integration, 9 known pitfalls from IMDA NSC production runs, corrected from failed full-SFT attempt

🎙️ CosyVoice LoRA 微调框架：LLM+Flow 联合训练，实现无 Prompt 语音合成

一个面向中文古诗词吟诵的 CosyVoice 1/2/3 LoRA 微调实验框架，系统比较三代架构在 text-to-speech token、Flow/DiT 声学建模、prompt 边界与吟诵风格迁移中的差异，并提供试听样例、客观评估指标和调试工具。

基于CosyVoice和ER-NeRF的语音克隆与视频生成

DaVinci Subtitle Alignment + Voice Clone + AI Emotion Optimization | CosyVoice2 TTS

Text-to-speech library for Rust with a unified trait interface over multiple backends (Kokoro ONNX). Part of the WaveKat voice pipeline.

一款基于 CosyVoice2 的《红楼梦》主题音色克隆体验平台。 该平台以网页形式呈现，专为展示与互动而设计，让用户通过沉浸式的语音克隆体验，感受 AI 的独特魅力。

阿里云百炼 CosyVoice 语音合成桌面应用（PySide6）+ DashScope 语音能力 requests 封装