Veed Fabric 1.0 is a specialized image-to-video model for high-fidelity lip synchronization (lip-sync). It generates precise speech animation and natural facial expressions over broad scene motion. This makes it particularly effective for talking avatar creation and educational content where realistic mouth movements are a primary requirement. This level of precision is achieved through a dual-input pipeline that processes visual and audio data simultaneously, mapping phonemes directly to facial keypoints. Videos can be produced in 480p or 720p.