AI Talking Head Generator

SadTalker
Create Realistic Talking Head Videos from One Image and Audio

SadTalker is an open-source neural talking head generation model. It creates expressive, lip-synced videos from a single face image and an audio track. The model predicts head motion, facial expression, and lip movement in sync with the audio, producing lifelike speaking animations. It supports full-head and upper-body talking styles with controllable pose and motion intensity.