AutoWebWorld: Synthesizing Infinite Verifiable Web Environments via Finite State Machines

2026-02-15

AI Summary

This work addresses the challenges of training autonomous Web GUI agents, which typically rely on costly and hard-to-verify real-world interaction data, further complicated by implicit state transitions that hinder validation. The authors propose modeling web environments as finite state machines (FSMs) and introduce a code-generation agent to automatically construct synthetic websites with explicitly defined states, actions, and transition rules. This enables fully automated trajectory generation and programmatic correctness verification. Using this approach, the method generates 11,663 verified trajectories at a cost of $0.04 per trajectory. A 7B-parameter model trained on this data outperforms all existing baselines on WebVoyager within 15 steps and demonstrates strong scaling behavior with increased data volume.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Web data generation and simulation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The performance of autonomous Web GUI agents heavily relies on the quality and quantity of their training data. However, a fundamental bottleneck persists: collecting interaction trajectories from real-world websites is expensive and difficult to verify. The underlying state transitions are hidden, leading to reliance on inconsistent and costly external verifiers to evaluate step-level correctness. To address this, we propose AutoWebWorld, a novel framework for synthesizing controllable and verifiable web environments by modeling them as Finite State Machines (FSMs) and use coding agents to translate FSMs into interactive websites. Unlike real websites, where state transitions are implicit, AutoWebWorld explicitly defines all states, actions, and transition rules. This enables programmatic verification: action correctness is checked against predefined rules, and task success is confirmed by reaching a goal state in the FSM graph. AutoWebWorld enables a fully automated search-and-verify pipeline, generating over 11,663 verified trajectories from 29 diverse web environments at only $0.04 per trajectory. Training on this synthetic data significantly boosts real-world performance. Our 7B Web GUI agent outperforms all baselines within 15 steps on WebVoyager. Furthermore, we observe a clear scaling law: as the synthetic data volume increases, performance on WebVoyager and Online-Mind2Web consistently improves.

Problem

• Web GUI agents
• training data
• state transitions
• trajectory verification
• autonomous agents

Innovation

• Finite State Machines
• Synthetic Web Environments
• Programmatic Verification
• Autonomous Web Agents
• Scalable Training Data