🎙️ Can AI preserve your voice across languages? See how VoxCPM2 transforms a video into a natural AI dub.

A developer built an open-source Video Dubbing Translator, enabling users to translate videos into different languages while preserving the original speaker’s voice and the natural flow of the conversation.

Powered by VoxCPM2, this project combines voice cloning, speech recognition, translation, and video reconstruction into a complete AI dubbing workflow:

✨ Reference-voice cloning with VoxCPM2 to preserve speaker identity ✨ Whisper for accurate transcription and word-level alignment ✨ Automatic translation for multilingual content creation ✨ Optional lip-sync support with LatentSync ✨ Local-first processing for more flexible and private workflows

From separating vocals and background audio to generating translated speech and rebuilding the final video, the project provides an end-to-end pipeline for creating AI-dubbed content.

🔗 Explore it: http://github.com/kadirb4rut/video-dubbing-translator

🤗 Model: http://huggingface.co/openbmb/VoxCPM2

🎬 Turn on sound to compare the original clip and the VoxCPM2 dubbing result.👇