Generative video · 2025
Sonic Lipsync AI
A cloud Gradio app that lip-syncs video from audio using the Sonic model — no local GPU theatre required.
Challenge
Running modern lipsync models locally is a non-starter for most collaborators.
Approach
I packaged Sonic behind a Colab-hosted Gradio interface with Hugging Face weights so anyone can generate from a browser.
Outcome
A practical generative-video tool that others can actually run. The public repo remains one of my most-starred experiments.
PythonGradioHugging FaceColab