StoryDiffusion (https://huggingface.co/spaces/YupengZhou/StoryDiffusion and documentation: https://github.com/HVision-NKU/StoryDiffusion) generates comic book or photo novel-style boxes from a list of instructions.

Generated from a character description or reference images and a box statement. Each line in the “Comic Description” field corresponds to a box.

Several styles are available and an automatic layout with classic comic book or four-box text bubbles. Advanced settings available: intensity of shared attention, number of reference images, dimensions, number of generation steps.

Based on Stable Diffusion’s image models.

Note: the text is not always well managed.

Open source and free, unlimited use at Hugging Face waiting times.