BookWorld: From Novels to Interactive Agent Societies for Creative Story Generation

TL;DR

BookWorld presents a book-grounded, scene-based multi-agent framework that constructs and simulates character societies from source books, orchestrated by role agents and a world agent. The pipeline integrates data extraction, worldbuilding, and retrieval-augmented memory, followed by post hoc rephrasing to produce novel-style narratives. Empirical results show BookWorld achieves high narrative quality and fidelity to source materials, outperforming baselines on several metrics with a reported win rate of 75.36% in certain evaluations. The work demonstrates scalable, artifact-rich story generation and interactive applications while acknowledging limitations in generalization and ethical considerations for content and use.

Abstract

Recent advances in large language models (LLMs) have enabled social simulation through multi-agent systems. Prior efforts focus on agent societies created from scratch, assigning agents with newly defined personas. However, simulating established fictional worlds and characters remain largely underexplored, despite its significant practical value. In this paper, we introduce BookWorld, a comprehensive system for constructing and simulating book-based multi-agent societies. BookWorld's design covers comprehensive real-world intricacies, including diverse and dynamic characters, fictional worldviews, geographical constraints and changes, e.t.c. BookWorld enables diverse applications including story generation, interactive games and social simulation, offering novel ways to extend and explore beloved fictional works. Through extensive experiments, we demonstrate that BookWorld generates creative, high-quality stories while maintaining fidelity to the source books, surpassing previous methods with a win rate of 75.36%. The code of this paper can be found at the project page: https://bookworld2025.github.io/.

Figures (4)

• Figure 1: A preview of BookWorld. Characters in the same location can interact with each other based on their goals and other factors. The map is represented as a discrete graph, where characters can traverse between adjacent locations spending several turns.
• Figure 2: The complete pipeline of BookWorld simulation. The framework starts from initializing the agents, then the agents interact across scenes.
• Figure 3: An example of role action $d_i$. Thoughts are enclosed in [ ]. Actions are enclosed in ( ). The character's dialogue is presented without any framing.
• Figure 4: The procedure of building worldview data. We extract atomic facts from each chunk, then filter, cluster and summarize them to construct the final dataset.