Qwen3-TTS - мощная модель для клонирования и генерации голоса локально на твоем компе.

Qwen3‑TTS is a sophisticated text‑to‑speech system developed by Alibaba’s Qwen, designed to generate natural, multilingual, real‑time audio. The video opens with an introduction at 00:00, then explains what Qwen3‑TTS is at 01:11, discusses who the model is for at 02:40, outlines the video’s purpose and plan at 03:10, details the custom nodes that will be used at 04:03, and shows how to support the channel at 04:27. It then demonstrates how to install custom ComfyUI‑QwenTTS nodes at 05:36, loads the models at 07:09, clones voices at 09:07, and manages default voices and emotions at 14:10. The design of voices is covered at 17:48, followed by a second custom ComfyUI‑Qwen‑TTS node and workflow at 18:49. The video explains how to organize model folders at 24:11, addresses a node installation error at 25:24, and concludes with pros and cons at 26:42. Finally, it shows how to prompt the instructor for Qwen3‑TTS at 30:22, discusses a block for continuous viewers at 33:05, and wraps up with a post‑script at 34:26.