Text to video
Text to video
A clip from words alone. On this desk that path is Kling 可灵, and it is not taking jobs yet.
Text to video means the model invents the frame from a description. Kling, also searched as Kling AI and 可灵视频, is the model reserved for that. It is not a switch inside MiniMax H3 or Seedance.
Until that runner is connected, every send on the desk is image to video: one photo plus an effect. The effect wording stays on the server.