StoryState: Agent-Based State Control for Consistent and Editable Storybooks

TL;DR

StoryState introduces an agent-based system that explicitly manages story state for multi-page storybook generation, enabling more consistent, editable, and localized edits across pages using prompt-based control.

Contribution

It presents a novel agent-based orchestration layer that explicitly models story state, improving multi-page story editing and consistency without retraining models.

Findings

Enables localized page edits with minimal unintended changes.

Improves cross-page visual consistency.

Reduces editing time compared to prior methods.

Abstract

Large multimodal models have enabled one-click storybook generation, where users provide a short description and receive a multi-page illustrated story. However, the underlying story state, such as characters, world settings, and page-level objects, remains implicit, making edits coarse-grained and often breaking visual consistency. We present StoryState, an agent-based orchestration layer that introduces an explicit and editable story state on top of training-free text-to-image generation. StoryState represents each story as a structured object composed of a character sheet, global settings, and per-page scene constraints, and employs a small set of LLM agents to maintain this state and derive 1Prompt1Story-style prompts for generation and editing. Operating purely through prompts, StoryState is model-agnostic and compatible with diverse generation backends. System-level experiments on multi-page editing tasks show that StoryState enables localized page edits, improves cross-page consistency, and reduces unintended changes, interaction turns, and editing time compared to 1Prompt1Story, while approaching the one-shot consistency of Gemini Storybook. Code is available at https://github.com/YuZhenyuLindy/StoryState