← Back to Skills Marketplace

Qwen3-TTS Voice Synthesis

Name: Qwen3-TTS Voice Synthesis
Author: vincentlau2046-sudo

by vincentlau2046-sudo · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

Downloads

Stars

Active Installs

Versions

Install in OpenClaw

/install tts-qwen3

Description

Qwen3-TTS 本地语音合成。支持音色克隆、声音设计、多角色对话。琪琪OPC首选TTS，失败回退到 tts-cosyvoice (Edge TTS)。

README (SKILL.md)

Qwen3-TTS 本地语音合成

基于 Qwen3-TTS 1.7B 模型，支持音色克隆 + 声音设计 + 多角色对话。 琪琪OPC 首选 TTS，失败时回退到 tts-cosyvoice (Edge TTS)。

优先级

Qwen3-TTS（本地GPU，音色克隆+设计）→ Edge TTS（tts-cosyvoice，云端回退）

琪琪OPC 音色库

6 个角色音色，通过 ComfyUI API 调用：

角色	音色名	方式	说明	用途
🐰 琪琪	qiqi_clone	克隆	ref_audio=qiqi_voice_v3.wav	琪琪对话
📖 旁白	narrator_teacher	VoiceDesign	seed=100, 温暖女声	叙事
👦 男孩	boy_child	VoiceDesign	seed=200, 活泼8岁	儿童男角
👧 女孩	girl_child	VoiceDesign	seed=300, 甜美7岁	儿童女角
👨 大人男	adult_male	VoiceDesign	seed=400, 沉稳	成年男角
👩 大人女	adult_female	VoiceDesign	seed=500, 优雅	成年女角

脚本

单角色 TTS

python3 {baseDir}/scripts/qwen_tts.py \
  --text "你好，我是琪琪" \
  --voice qiqi_clone \
  --output /tmp/output.wav

多角色对话 TTS

python3 {baseDir}/scripts/qwen_tts_dialogue.py \
  --script "琪琪:你好呀！\
旁白:琪琪开心地笑了。" \
  --output /tmp/dialogue.wav \
  --srt /tmp/dialogue.srt

选项

选项	默认	说明
`--text`	(必需)	要合成的文本
`--voice`	narrator_teacher	音色名（见音色库）
`--output`	/tmp/qwen_tts_output.wav	输出文件路径
`--language`	Chinese	语言
`--model`	1.7B	模型大小 (0.6B/1.7B)
`--attention`	sdpa	注意力机制
`--fallback-edge`	true	失败时回退到 Edge TTS

对话脚本格式

角色名: 台词内容
角色名: 台词内容

角色名映射到音色库中的音色。旁白 映射到 narrator_teacher。

依赖

ComfyUI 运行中（localhost:8188）
ComfyUI-Qwen-TTS 插件已安装
Qwen3-TTS 模型已下载（~/ComfyUI/models/qwen-tts/）
comfyui-venv Python 环境

回退策略

当 Qwen3-TTS 不可用时（ComfyUI 未启动 / GPU 显存不足 / 生成失败），自动回退到 Edge TTS (tts-cosyvoice)：

琪琪 → zh-CN-XiaoyiNeural
旁白 → zh-CN-XiaoxiaoNeural
男孩 → zh-CN-YunxiNeural
女孩 → zh-CN-XiaoyiNeural
大人男 → zh-CN-YunjianNeural
大人女 → zh-CN-XiaoxiaoNeural

版本: v1.0 | 琪琪OPC 首选 TTS | 基于 Qwen3-TTS + ComfyUI

Usage Guidance

Install only if you are comfortable with possible off-device text processing. Keep COMFYUI_URL set to localhost or another trusted ComfyUI server, avoid sensitive text unless you disable fallback with --fallback-edge false, and review or install the referenced tts-cosyvoice dependency before relying on fallback output.

Capability Assessment

ℹ Purpose & Capability

The ComfyUI Qwen3-TTS workflow, voice presets, audio conversion, and output writing fit the stated voice-synthesis purpose, but the documented dialogue helper is absent and the local-TTS framing is weakened by automatic fallback behavior.

⚠ Instruction Scope

SKILL.md discloses Edge TTS fallback and the fallback flag, but it does not prominently warn that user text may be sent off-device when local generation fails; it also omits the --file input and ffmpeg behavior from the main capability description.

ℹ Install Mechanism

No installer, package download, persistence hook, or hidden setup script is present, but the runtime depends on local ComfyUI assets and a separate tts-cosyvoice script that are not declared as formal requirements.

⚠ Credentials

Network access is expected for ComfyUI, but COMFYUI_URL is accepted from the environment without host validation, so generated text and workflow data could be sent to a non-local endpoint if that variable is changed.

ℹ Persistence & Privilege

No autostart, privilege escalation, credential harvesting, or broad indexing was found; file mutation is limited to user output paths and copying reference audio into the ComfyUI input directory for the voice-clone flow.

How to Use

Make sure OpenClaw is installed (local or Docker)
Run the install command in chat: /install tts-qwen3
After installation, invoke the skill by name or use /tts-qwen3
Provide required inputs per the skill's parameter spec and get structured output

Version History

v1.0.0

Qwen3-TTS 本地语音合成，支持音色克隆+声音设计+6角色音色库，失败回退Edge TTS

Metadata

Slug tts-qwen3

Version 1.0.0

License MIT-0

All-time Installs 0

Active Installs 0

Total Versions 1

Frequently Asked Questions

What is Qwen3-TTS Voice Synthesis?

Qwen3-TTS 本地语音合成。支持音色克隆、声音设计、多角色对话。琪琪OPC首选TTS，失败回退到 tts-cosyvoice (Edge TTS)。 It is an AI Agent Skill for Claude Code / OpenClaw, with 40 downloads so far.

How do I install Qwen3-TTS Voice Synthesis?

Run "/install tts-qwen3" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Qwen3-TTS Voice Synthesis free?

Yes, Qwen3-TTS Voice Synthesis is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Qwen3-TTS Voice Synthesis support?

Qwen3-TTS Voice Synthesis is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Qwen3-TTS Voice Synthesis?

It is built and maintained by vincentlau2046-sudo (@vincentlau2046-sudo); the current version is v1.0.0.

More Skills