Extend a video based on the context of current generated video frames, input prompt

Hi, I am currently developing a video generation system based on the current pipelines. I would appreciate if there is also video extension-based pipeline where we can provide the current video frames as context and the input prompt as a direction for generating the next set of frames. If there exists such feature, please let me know as I am not aware of it.

Thanks!