no lipsync on my ai-generated character

lipsync-2 learns from the input mouth movements so it gives best results when input video has good mouth movements. So we recommend users to add "generate character with natural mouth movements" to their prompt when they are generating the video using any AI video generation tool

Additional Technical Considerations for Lipsync Success

Model Performance by Content Type: The lipsync model works best with photo‑realistic human videos and may struggle with static CGI or cartoon characters. If you're working with animated or heavily stylized content, expect reduced accuracy.

Input Requirements: The model requires visible lip movements in your selected segment to function properly. Even with good audio, segments with minimal or no visible mouth movement will produce poor results regardless of prompt optimization.

Important limitation: Lipsync will not work during video segments that contain still frames where the speaker is not actively moving or speaking, even if audio is present. This can cause lips to stop moving mid-sentence in some cases. Ensure your input video maintains consistent natural speaking motion throughout to avoid intermittent lipsync failures and to get the best results.