WebNavigator: Global Web Navigation via Interaction Graph Retrieval

2026-03-20

AI Summary

Existing autonomous web navigation methods suffer from “topological blindness,” relying heavily on trial-and-error exploration and struggling to achieve efficient navigation in complex scenarios. This work proposes a novel interaction-graph-based “retrieve–reason–teleport” framework that, for the first time, leverages zero-token-cost offline heuristic exploration to construct a global topological map. This paradigm shift transforms navigation from probabilistic exploration to deterministic path planning. The proposed method achieves state-of-the-art performance on WebArena and OnlineMind2Web benchmarks, attaining a 72.9% success rate on multi-site WebArena tasks—more than double the performance of current enterprise-grade agents.

Technology Category

• Planning, Routing, and Scheduling: Model-Based Reasoning
• Search and Optimization: Heuristic Search
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Despite significant advances in autonomous web navigation, current methods remain far from human-level performance in complex web environments. We argue that this limitation stems from Topological Blindness, where agents are forced to explore via trial-and-error without access to the global topological structure of the environment. To overcome this limitation, we introduce WebNavigator, which reframes web navigation from probabilistic exploration into deterministic retrieval and pathfinding. WebNavigator constructs Interaction Graphs via zero-token cost heuristic exploration offline and implements a Retrieve-Reason-Teleport workflow for global navigation online. WebNavigator achieves state-of-the-art performance on WebArena and OnlineMind2Web. On WebArena multi-site tasks, WebNavigator achieves a 72.9\% success rate, more than doubling the performance of enterprise-level agents. This work reveals that Topological Blindness, rather than model reasoning capabilities alone, is an underestimated bottleneck in autonomous web navigation.

Problem

• autonomous web navigation
• Topological Blindness
• global topological structure
• web environment
• navigation bottleneck

Innovation

• Interaction Graph
• Topological Blindness
• Global Navigation
• Retrieve-Reason-Teleport
• Zero-token Exploration