Skip to content
SPECIFICTION
Get listed

KeySync

antonibigata.github.io

Web
Visit antonibigata.github.io
Spec score 7.3 #22 of 28 in AI Video Lip Sync Tools

Spec sheet

Purpose
KeySync synchronizes lip movements in an existing video with new input audio.Read 9 Oct 2026 · antonibigata.github.io
License
The GitHub repository identifies its license as Apache-2.0.Read 9 Oct 2026 · github.com
Architecture
KeySync is a two-stage framework designed to improve temporal consistency in lip synchronization.Read 9 Oct 2026 · antonibigata.github.io
Demo limit
The Hugging Face demo is limited to videos of up to 6 seconds because of GPU restrictions.Read 9 Oct 2026 · github.com
Evaluation
The repository includes a LipScore metric for measuring synchronization quality between generated and ground-truth videos.Read 9 Oct 2026 · github.com
Expression leakage
Its masking strategy is designed to reduce expression leakage from the input video.Read 4 Oct 2026 · antonibigata.github.io
Framework
KeySync uses a two-stage framework to improve temporal consistency in lip synchronization.Read 4 Oct 2026 · antonibigata.github.io
Input formats
The quick start guide uses MP4 video files and WAV audio files as inputs.Read 4 Oct 2026 · github.com
Input preparation
The repository recommends converting video to 25 fps and audio to 16000 Hz for compatibility with the models.Read 9 Oct 2026 · github.com
Interactive demo
The demo accepts video and audio uploads to create synchronized videos and is limited to videos up to 6 seconds by Hugging Face Spaces GPU restrictions.Read 4 Oct 2026 · github.com
Leakage and occlusion handling
The framework uses a masking strategy to address expression leakage and facial occlusions.Read 9 Oct 2026 · antonibigata.github.io
Local requirements
The repository lists a CUDA-compatible GPU, Python 3.11, and Conda as prerequisites for local use.Read 9 Oct 2026 · github.com
Model weights
The repository instructs users to download pretrained models from the toninio19/keysync Hugging Face repository.Read 4 Oct 2026 · github.com
Occlusion controls
The command-line options include a setting to enable occlusion handling and coordinates for the object to mask.Read 4 Oct 2026 · github.com
Occlusion option
The inference parameters include an option to mask objects that block the face.Read 9 Oct 2026 · github.com
Occlusions
The framework includes a masking approach for handling facial occlusions.Read 4 Oct 2026 · antonibigata.github.io
Online demo
The interactive demo accepts video and audio uploads to create synchronized videos.Read 9 Oct 2026 · github.com
Preprocessing
Provided scripts can convert videos to 25 fps and audio to 16000 Hz, and crop videos around the face.Read 4 Oct 2026 · github.com
Pretrained models
The repository provides instructions to download pretrained models from Hugging Face.Read 9 Oct 2026 · github.com
Reported results
The authors report improved lip reconstruction and cross-synchronization, and reduced expression leakage measured with their LipLeak metric.Read 9 Oct 2026 · antonibigata.github.io
Training
The repository includes separate scripts for training the keyframe and interpolation models.Read 9 Oct 2026 · github.com

Where Specifiction places it

Alternatives to KeySyncAll alternatives →