voice-pro: an all-in-one AI dubbing studio for YouTube processing, voice cloning, and multilingual translation

What it solves

Voice-Pro is an all-in-one multimedia processing tool designed to simplify the complex workflow of creating multilingual content. It eliminates the need to jump between separate tools for downloading videos, separating audio, transcribing speech, translating text, and generating new voice-overs.

How it works

The application provides a Gradio-based web interface that integrates several state-of-the-art AI models:

• Speech-to-Text (ASR): Uses Whisper, Faster-Whisper, and Whisper-Timestamped for high-accuracy transcription and subtitle generation.
• Text-to-Speech (TTS): Employs Edge-TTS, kokoro, and zero-shot cloning models like F5-TTS, E2-TTS, and CosyVoice to generate natural-sounding speech or clone specific voices.
• Translation: Integrates Deep-Translator (and optionally Azure Translator) to support over 100 languages.
• Audio Processing: Uses Demucs for voice separation and yt-dlp for YouTube content extraction.

Who it’s for

This tool is primarily aimed at content creators, podcasters, researchers, and multilingual professionals who need to dub videos, create subtitles, or produce voice-cloned podcasts.

Highlights

• Complete Dubbing Pipeline: Integrated YouTube downloading, noise removal, transcription, translation, and TTS in one studio.
• Zero-Shot Voice Cloning: Ability to clone voices using F5-TTS, E2-TTS, and CosyVoice without extensive training.
• Multilingual Support: Speech recognition and translation capabilities for over 100 languages.
• Real-Time Capabilities: Supports instant speech recognition and on-the-fly translation.
• User-Friendly Installation: Uses uv for fast, reproducible installs on Windows with NVIDIA GPU support.