This is a simple workflow that takes each frame from a video, passes it through img2img (with loras and optional autocaption), then recombines them into a new video.
This is my first published workflow, so feel free to comment/criticize/help :D
Frame-by-Frame img2img (autocap).json.zip