Measure a library of images against your taxonomy & the Canon world.
Pick any combination. A single model call per image runs all selected modes at once.
No saved characters yet — add some on the Characters tab.
The text profile is built automatically from the reference images using the model selected on the Classify tab, and is revised whenever you add or remove images. References may be a single image, a contact-sheet file holding several images, or many files.
Describe the appearance & background. You'll get a profile image to approve, then a bio to approve; then the card, gown, and full bio sheet are generated automatically (a random example sheet sets the style).
Every character is a bio sheet in character/sheets/ — portrait, Prompt, reference images and cards. Click one to open it.
Loading…
A reusable image-generation request: pick characters & attributes from your known sets, then Generate now to create images (auto-saved to this scene's library).
Every analyzed image, across all runs. Filter, sort, select, and export.
Loading…
No runs recorded.
Generated images currently in the library, by scene and by character. An image with multiple characters counts toward each, so per-character costs can sum above the total.
Paste a run-on text of saved prompts. OpenAI splits it into individual prompts and
classifies each as a scene, a universal effect, or a scene + effect.
Lines of ==== mark sections whose header applies to the prompts beneath. Then tick the
ones to save as Prompts.