English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

BrowserForge: Scaling Web Agent Training via Parallel Browser Sandboxes on the Open Web

Forum topic · 小凯 · 2026-08-27

Summary

BrowserForge is a research framework for generating large-scale web interaction data to train pixel-based web agents. Acting from rendered pixels avoids the fragility and high token costs of parsing HTML or accessibility trees, but training such agents requires massive high-quality trajectories that existing datasets and synthesis pipelines fail to provide—they cover only a few thousand trajectories from narrow, fixed website lists. BrowserForge addresses this by driving many browser sandboxes in parallel across the open web through three coupled components: an open-web sourcing stage exposing agents to hundreds of thousands of real, publicly reachable websites; a sandbox cluster manager scheduling hundreds of concurrent browsers with high utilization; and a proposer-solver dual-agent loop that converts raw pages into executable tasks and collects verified trajectories. Paper by Fei Tang et al., arXiv:2608.24848.

Overview

Field: NLP Authors: Fei Tang, Huawen Shen, Zhiqiong Lu, Zhengxi Lu, Pengyuan Lyu, et al. Published: 2026-08-25 arXiv: 2608.24848

Abstract (translated from Chinese summary)

Web agents that act from rendered pixels avoid the fragility and heavy token cost of reading a page's HTML or accessibility tree, but training them depends on large amounts of high-quality interaction trajectories, and how to produce such data at scale remains an open problem. Public datasets typically contain only a few thousand trajectories drawn from a fixed and narrow set of websites, and even recent automated synthesis pipelines stay bound to predefined site lists or tutorial sources, so the number of distinct websites the agent ever sees barely grows.

The authors present BrowserForge, a framework that generates web interaction data at scale by driving many browser sandboxes in parallel over the open web. BrowserForge couples three components:

1. Open-web sourcing stage — exposes the agent to hundreds of thousands of real, publicly reachable websites. 2. Sandbox cluster manager — schedules hundreds of concurrent browsers and achieves high utilization. 3. Proposer-solver dual-agent loop — converts raw pages into executable tasks, then collects verified trajectories.

Original Abstract (excerpt)

> Web agents that act from rendered pixels avoid the fragility and heavy token cost of reading a page's HTML or accessibility tree, but training them depends on large amounts of high-quality interaction trajectories, and how to produce such data at scale remains an open problem. Public datasets typically contain only a few thousand trajectories drawn from a fixed and narrow set of websites... We present BrowserForge, a framework that generates web interaction data at scale by driving many browser sandboxes in parallel over the open web.

---

*Auto-collected on 2026-08-27*

Tags

#web-agents#browserforge#nlp#arxiv#data-synthesis#browser-automation#machine-learning

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634090