Mango: Multi-Agent Web Navigation via Global-View Optimization

TL;DR

Mango is a multi-agent web navigation method that uses website structure and global view optimization to improve success rates in reaching target information, outperforming baselines.

Contribution

It introduces a novel approach combining global view optimization, multi-armed bandit URL selection, and episodic memory for efficient web navigation.

Findings

Achieves 63.6% success rate on WebVoyager, outperforming baseline by 7.3%.

Attains 52.5% success rate on WebWalkerQA, surpassing baseline by 26.8%.

Demonstrates generalizability across different model backbones.

Abstract

Existing web agents typically initiate exploration from the root URL, which is inefficient for complex websites with deep hierarchical structures. Without a global view of the website's structure, agents frequently fall into navigation traps, explore irrelevant branches, or fail to reach target information within a limited budget. We propose Mango, a multi-agent web navigation method that leverages the website structure to dynamically determine optimal starting points. We formulate URL selection as a multi-armed bandit problem and employ Thompson Sampling to adaptively allocate the navigation budget across candidate URLs. Furthermore, we introduce an episodic memory component to store navigation history, enabling the agent to learn from previous attempts. Experiments on WebVoyager demonstrate that Mango achieves a success rate of 63.6% when using GPT-5-mini, outperforming the best baseline by 7.3%. Furthermore, on WebWalkerQA, Mango attains a 52.5% success rate, surpassing the best baseline by 26.8%. We also demonstrate the generalizability of Mango using both open-source and closed-source models as backbones. Our data and code are open-source and available at https://github.com/VichyTong/Mango.