Analyze a video and detect all human faces in it. Returns a session_id and a face_list that are required for the subsequent lip sync generation step.
This is a synchronous call — the result is returned immediately, not as an async task.
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Public URL of the video to analyze. The video must contain at least one clearly visible human face.
Required if video_id is not provided.
"https://example.com/person-speaking.mp4"
ID of a video previously generated by Kling (e.g. from a text-to-video or image-to-video task result).
Required if video_url is not provided.
"860260754053148756"