Qwen3 Tts Skills

This is a complete local TTS workflow built around Qwen3-TTS models with three distinct modes: CustomVoice for built-in speakers with emotion control, VoiceDesign for describing voices in natural language (like "high-pitched loli voice"), and VoiceClone for mimicking reference audio. The real utility is in batch dubbing long articles into multi-voice audio with automatic speaker assignment and emotion tagging, then merging everything with FFmpeg. Supports Chinese, English, Japanese, and Korean with speakers like Vivian and Ryan out of the box. The documentation is thorough with actual command examples, though it's heavy on Chinese content. You'll need a GPU for reasonable performance and FFmpeg installed for the batch features. If you're doing voiceovers for articles, audiobooks, or multi-character dialogue, this handles the whole pipeline from text splitting to final WAV.

skills.sh installs: 141
GitHub stars: 13
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add mu-zi-lee/qwen3-tts-skill --skill qwen3-tts-skills --agent claude-code

Installs into .claude/skills of the current project.

Files

Qwen3-TTS 技能

将文本转换为高质量语音的完整工作流。

🚀 快速开始

场景 1：单句语音生成

直接调用脚本生成语音：

# 中文语音（默认 Vivian 女声）
uv run qwen3-tts-skills/scripts/run_qwen3_tts.py custom-voice \
  --language Chinese \
  --text "你好，欢迎使用语音合成。" \
  --out-dir outputs

# 英文语音（默认 Ryan 男声）
uv run qwen3-tts-skills/scripts/run_qwen3_tts.py custom-voice \
  --language English \
  --text "Hello, welcome to text-to-speech." \
  --out-dir outputs

场景 2：长文稿批量配音

将文章转换为完整语音文件：

用户文稿 → [AI分析生成配音稿] → [用户审核] → [批量TTS] → 完整语音.wav

详见下方 长文稿批量配音 章节。

📋 模型选择指南

根据需求选择合适的模型：

模式	模型	适用场景	命令
CustomVoice	`Qwen3-TTS-12Hz-1.7B-CustomVoice`	使用内置音色 + 情感控制	`custom-voice`
VoiceDesign	`Qwen3-TTS-12Hz-1.7B-VoiceDesign`	用自然语言描述想要的音色	`voice-design`
VoiceClone	`Qwen3-TTS-12Hz-1.7B-Base`	克隆参考音频的声音	`voice-clone`
Tokenizer	`Qwen3-TTS-Tokenizer-12Hz`	音频编解码	`tokenizer-roundtrip`