Research Demo

Autonomous
Video Dubbing
with MuseTalk

Lip-sync any talking-head video to a new audio track in seconds — powered by MuseTalk v1.5 running on a single Kaggle T4 GPU.

Side-by-side demos

How it works

Option A — Generate audio with TTS
📝
Text input
MOSS-TTS-Nano
→
🔊
Audio .wav
16 kHz mono
→
🎬
Target video
512 × 512
→
⚡
MuseTalk v1.5
Kaggle T4 GPU
→
✅
Dubbed video
lip-synced .mp4
or
Option B — Paste your own audio directly
🎤
Your audio file
.wav / .mp3
→
🎬
Target video
512 × 512
→
⚡
MuseTalk v1.5
Kaggle T4 GPU
→
✅
Dubbed video
lip-synced .mp4