I developed my initial idea with Perplexity AI, iterating on various concepts until the story felt coherent, I also used Perplexity AI to generate a series of shot descriptions that I could then use as prompts in Midjourney.

For the overall style, I created a Perplexity AI space and fed it creative direction that would consistently describe the aesthetic of the scene to help maintain a visual coherence and consistency between shot prompts.

I used a number of image generations within Midjourney to develop character sheets and then re-used those character sheets as references with each shot generation.

Each image was then upscaled and used as image-to-video references in both Luma AI and Kling AI to generate the videos before re-upscaling to 4K with Topaz Labs.

The audio was cut together from multiple iterations within Udio, I was keen to explore how character themes could be created using generative AI audio.

The final colour grading, edit and composition was made with Capcut.