Create Lip Sync Task

Create an asynchronous lip sync video generation task. The facial movements in the source video are re-animated to match the provided speech.

Prerequisite: call the identify-face endpoint first to obtain a session_id.

Two input modes are supported:

• Text mode — provide text, voice_id, and voice_language. The platform converts the text to speech using the specified voice and drives the lip movements.
• Audio mode — provide audio_url. The lip movements are driven directly by the supplied audio file.

After submission, poll GET /kling/v1/videos/advanced-lip-sync/{task_id} until status is succeeded.

{
  "input": {
    "session_id": "abc123-session-id",
    "face_image_url": "https://example.com/face.jpg",
    "text": "你好，欢迎来到我的频道",
    "voice_id": "girlfriend_1_cn",
    "voice_language": "zh",
    "audio_url": "https://example.com/speech.mp3"
  }
}

{
  "id": "task_abc123",
  "task_id": "task_abc123",
  "object": "video",
  "model": "kling-lip-sync",
  "created_at": 1773812605
}