easyvoice

Project Overview

EasyVoice is a local multilingual傻瓜 dubbing / voice cloning tool based on Qwen3-TTS (Chinese interface, ready to use out of the box).

Key Features

• 3-second voice cloning: Upload or record about 3 seconds of reference audio to clone that voice for any text/subtitle dubbing
• Multilingual dubbing: Supports Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian (10 languages), defaulting to auto language detection
• Subtitle dubbing (multi-role): Upload SRT/VTT/LRC subtitles, automatically assign speaker voices based on "Name:" prefixes, preview/edit/re-record line by line, generate aligned to the timeline
• Local operation: Full local inference, no cloud upload
• GPU auto-adaptation: Automatically uses CUDA acceleration with a GPU, falls back to CPU without one
• Quality options: Manually switch between "Fast 0.6B / High Quality 1.7B" models
• Out-of-the-box: Download the package, extract, and double-click to run

Target Audience

• Users who need dubbing for videos or subtitles
• Users who need to clone specific voices
• Video creators and self-media operators

Quick Start

1. Go to Releases to download the package (full ~4.6GB or CPU slim ~0.5GB)
2. Extract to a local directory
3. Double-click the startup script to run
4. Upload reference audio to clone a voice, enter text or upload subtitles for dubbing

Core Advantages

• 3-second ultra-fast voice cloning with simple operation
• Supports 10 languages covering major languages
• Local inference for privacy and security
• GPU auto-adaptation; works without a GPU too
• Chinese subtitle multi-role dubbing aligned to timeline

System Requirements

• OS: Windows
• Additional dependencies: Full package recommended with NVIDIA GPU; CPU slim package for systems without a GPU