story-cover

To generate a web novel cover, provide the book title and author name so the system can infer the genre. It then builds layered prompts and calls GPT-Image-2 to output a professional cover image with text.

Yes, you can use reference images for AI book cover design. The workflow supports both text-to-image and image-to-image generation, allowing you to provide optional reference imagery to guide the GPT-Image-2 output.

Yes, the generated book cover aligns with specific web novel platform styles. The system applies visual aesthetics from references/cover-styles.md to iterate across platform-specific styles and produce multiple candidate covers.

Before running the novel cover generation workflow, you need to set the required environment variables and define a BOOK_DIR input. You also need to provide the book title and author name to start the automated process.

The system determines the visual style for your book cover by analyzing the provided title and author name. It uses this information to infer the genre and construct layered prompts that match the appropriate platform aesthetics.

The final AI generated book cover is saved as a base64 image file. Along with the image output, the system also saves the prompt text used for the GPT-Image-2 generation to support iteration.