Recursively writes descriptions of video scenes using Large Language Models and Image Captioners - View it on GitHub