Web World Models

TL;DR

The paper introduces Web World Models (WWMs), a hybrid approach combining web code for logical consistency with large language models for high-level generation, enabling scalable, controllable, and open-ended virtual worlds.

Contribution

It proposes a novel framework that integrates web programming with language models to create structured, scalable virtual environments with practical design principles.

Findings

WWMs enable structured yet open-ended worlds.

Web stacks can serve as scalable substrates for world models.

Design principles improve controllability and diversity of environments.

Abstract

Language agents increasingly require persistent worlds in which they can act, remember, and learn. Existing approaches sit at two extremes: conventional web frameworks provide reliable but fixed contexts backed by databases, while fully generative world models aim for unlimited environments at the expense of controllability and practical engineering. In this work, we introduce the Web World Model (WWM), a middle ground where world state and ``physics'' are implemented in ordinary web code to ensure logical consistency, while large language models generate context, narratives, and high-level decisions on top of this structured latent state. We build a suite of WWMs on a realistic web stack, including an infinite travel atlas grounded in real geography, fictional galaxy explorers, web-scale encyclopedic and narrative worlds, and simulation- and game-like environments. Across these systems, we identify practical design principles for WWMs: separating code-defined rules from model-driven imagination, representing latent state as typed web interfaces, and utilizing deterministic generation to achieve unlimited but structured exploration. Our results suggest that web stacks themselves can serve as a scalable substrate for world models, enabling controllable yet open-ended environments. Project Page: https://github.com/Princeton-AI2-Lab/Web-World-Models.