Elamur Research⌕
☰
home / papers / 2026 / WebDART: Dynamic Decomposition and Re-planning for Complex Web Tasks

WebDART: Dynamic Decomposition and Re-planning for Complex Web Tasks

Paper · 2026 · Jingbo Yang, Bairu Hou, Wei Wei et al.

Summary

LLM-based web agents handle short tasks like opening a page or filling a form. They break down on "chores" that combine long multi-page browsing, collecting many data points and reasoning over constraints. For example, a GPT-4o agent scores 8.0% on WebChoreArena but 46.6% on WebArena. The authors attribute this to overload: the model must browse, remember and compute all at once.

WebDART is a training-free framework for a single frozen LLM working over accessibility-tree text. It splits each objective into three stages handled in sequence:

  • navigation, to reach the relevant pages;
  • information extraction, to turn page content into structured records;
  • execution, to analyse the data, e.g., via Python, or act on it.

During navigation, the agent revises the plan step by step as new pages appear. This lets it exploit filters or sort options it discovers along the way.

Across three LLM backbones, WebDART clearly beats prior agents on WebChoreArena while staying competitive on WebArena. It also transfers to live websites. The work shows that structured decomposition can boost agents on complex tasks without extra training.

Key points

  • On WebChoreArena, WebDART beats BrowserGym and AgentOccam by up to 13.7 percentage points. Dynamic re-planning adds 7.7 points on shopping tasks and cuts navigation by 14.7 steps on average.
  • On WebArena, the decomposition stage is skipped. WebDART reaches 48.1% overall, compared with 46.6% for AgentOccam.
  • On WebVoyager (13 real websites, GPT-4o backbone), it scores 45.74% versus 41.09% for AgentOccam. It wins or ties on 11 of the 13 sites.
  • In an error analysis of 100 sampled trajectories, navigation causes nearly 60% of failures. Typical problems are lost context, stopping too early and drifting to irrelevant pages.
  • Limitations: the method is text-only (no screenshots) and re-planning is not learned. Code is available at github.com/UCSB-NLP-Chang/WebDART.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🏢AccentureZurich · Switzerland6 works · 27 citations🏢Accenture (United States)New York · United States5 works · 261 citations

 Fields, topics and concepts

📚 Shelves 3The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksInformation Systems · 1,437 works

Cite

Cite as Jingbo Yang, Bairu Hou, Wei Wei, Shiyu Chang and Yujia Bao. 2026. WebDART: Dynamic Decomposition and Re-planning for Complex Web Tasks. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.568

@inproceedings{yang2026webdart,
  title = {{WebDART: Dynamic Decomposition and Re-planning for Complex Web Tasks}},
  author = {Jingbo Yang, Bairu Hou, Wei Wei, Shiyu Chang and Yujia Bao},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.568},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.568}
}

References

27 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Alfonso Amayuelas, Jingbo Yang, Saaket Agashe, Ashwin Nagarajan, Antonis Antoniades, Xin Eric Wang, and William Wang. 2025. Self-resource allocation in multi-agent llm systems. arXiv preprint arXiv:2504.02051. — [arXiv:2504.02051](https://arxiv.org/abs/2504.02051) find
  2. De Chezelles, Thibault Le Sellier, Sahar Omidi Shayegan, Lawrence Keunho Jang, Xing Han Lù, Ori Yoran, Dehan Kong, Frank F Xu, Siva Reddy, Quentin Cappart, and 1 others. 2024. The browsergym ecosystem for web agent research. arXiv preprint arXiv:2412.05467. find
  3. Alexandre Drouin, Maxime Gasse, Massimo Caccia, Issam H Laradji, Manuel Del Verme, Tom Marty, David Vazquez, Nicolas Chapados, and Alexandre Lacoste. 2024. Workarena: How capable are web agents at solving common knowledge work tasks? In International Conference on Machine Learning, pages 11642–11662… find
  4. Hongliang He, Wenlin Yao, Kaixin Ma, Wenhao Yu, Yong Dai, Hongming Zhang, Zhenzhong Lan, and Dong Yu. 2024. Webvoyager: Building an end-to-end web agent with large multimodal models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), p…
  5. Wenyi Hong, Weihan Wang, Qingsong Lv, Jiazheng Xu, Wenmeng Yu, Junhui Ji, Yan Wang, Zihan Wang, Yuxiao Dong, Ming Ding, and 1 others. 2024. Cogagent: A visual language model for gui agents. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14281–14290.
  6. Jing Yu Koh, Robert Lo, Lawrence Jang, Vikram Duvvur, Ming Chong Lim, Po-Yu Huang, Graham Neubig, Shuyan Zhou, Ruslan Salakhutdinov, and Daniel Fried. 2024. Visualwebarena: Evaluating multimodal agents on realistic visual web tasks. arXiv preprint arXiv:2401.13649.
  7. Dongjun Lee, Juyong Lee, Kyuyoung Kim, Jihoon Tack, Jinwoo Shin, Yee Whye Teh, and Kimin Lee. 2025. Learning to contextualize web pages for enhanced decision making by llm agents. arXiv preprint arXiv:2503.10689. — [arXiv:2503.10689](https://arxiv.org/abs/2503.10689) find
  8. Alexander Li, Lerrel Pinto, and Pieter Abbeel. 2020. Generalized hindsight for reinforcement learning. Advances in neural information processing systems, 33:7754–7767. find
  9. Hongjin Su, Ruoxi Sun, Jinsung Yoon, Pengcheng Yin, Tao Yu, and Sercan Ö Arık. 2025. Learn-by-interact: A data-centric framework for self-adaptive agents in realistic environments. arXiv preprint arXiv:2501.10893. — [arXiv:2501.10893](https://arxiv.org/abs/2501.10893) find
  10. Zora Zhiruo Wang, Jiayuan Mao, Daniel Fried, and Graham Neubig. 2024. Agent workflow memory. arXiv preprint arXiv:2409.07429.
  11. Jason Wei, Zhiqing Sun, Spencer Papay, Scott McKinney, Jeffrey Han, Isa Fulford, Hyung Won Chung, Alex Tachard Passos, William Fedus, and Amelia Glaese. 2025. Browsecomp: A simple yet challenging benchmark for browsing agents. arXiv preprint arXiv:2504.12516. find
  12. Evan Zheran Liu, Kelvin Guu, Panupong Pasupat, Tianlin Shi, and Percy Liang. 2018. Reinforcement learning on web interfaces using workflow-guided exploration. arXiv preprint arXiv:1802.08802. find
  13. Ke Yang, Yao Liu, Sapana Chaudhary, Rasool Fakoor, Pratik Chaudhari, George Karypis, and Huzefa Rangwala. 2024a. Agentoccam: A simple yet strong baseline for llm-based web agents. arXiv preprint arXiv:2410.13825. — [arXiv:2410.13825](https://arxiv.org/abs/2410.13825) find
  14. Grégoire Mialon, Clémentine Fourrier, Thomas Wolf, Yann LeCun, and Thomas Scialom. 2023. Gaia: a benchmark for general ai assistants. In The Twelfth International Conference on Learning Representations.
  15. Yan Yang, Dongxu Li, Yutong Dai, Yuhao Yang, Ziyang Luo, Zirui Zhao, Zhiyuan Hu, Junzhe Huang, Amrita Saha, Zeyuan Chen, and 1 others. 2025. Gta1: Gui test-time scaling agent. arXiv preprint arXiv:2507.05791. — [arXiv:2507.05791](https://arxiv.org/abs/2507.05791) find
  16. Atsuyuki Miyai, Zaiying Zhao, Kazuki Egashira, Atsuki Sato, Tatsumi Sunada, Shota Onohara, Hiromasa Yamanishi, Mashiro Toyooka, Kunato Nishina, Ryoma Maeda, and 1 others. 2025. Webchorearena: Evaluating web browsing agents on realistic tedious web tasks. arXiv preprint arXiv:2506.01952. — [arXiv:2506.01952](https://arxiv.org/abs/2506.01952) find
  17. Yuhao Yang, Yue Wang, Dongxu Li, Ziyang Luo, Bei Chen, Chao Huang, and Junnan Li. 2024b. Aria-ui: Visual grounding for gui instructions. arXiv preprint arXiv:2412.16256. find
  18. Jiayi Pan, Yichi Zhang, Nicholas Tomlin, Yifei Zhou, Sergey Levine, and Alane Suhr. 2024. Autonomous evaluation and refinement of digital agents. arXiv preprint arXiv:2404.06474. find
  19. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2023. React: Synergizing reasoning and acting in language models. In International Conference on Learning Representations (ICLR).
  20. Zehan Qi, Xiao Liu, Iat Long Iong, Hanyu Lai, Xueqiao Sun, Wenyi Zhao, Yu Yang, Xinyue Yang, Jiadai Sun, Shuntian Yao, and 1 others. 2024. Webrl: Training llm web agents via self-evolving online curriculum reinforcement learning. arXiv preprint arXiv:2411.02337. — [arXiv:2411.02337](https://arxiv.org/abs/2411.02337) find
  21. Ruichen Zhang, Mufan Qiu, Zhen Tan, Mohan Zhang, Vincent Lu, Jie Peng, Kaidi Xu, Leandro Z Agudelo, Peter Qian, and Tianlong Chen. 2025a. Symbiotic cooperation for web agents: Harnessing complementary strengths of large and small llms. arXiv preprint arXiv:2502.07942. — [arXiv:2502.07942](https://arxiv.org/abs/2502.07942) find
  22. Yujia Qin, Yining Ye, Junjie Fang, Haoming Wang, Shihao Liang, Shizuo Tian, Junda Zhang, Jiahao Li, Yunxin Li, Shijue Huang, and 1 others. Ui-tars: Pioneering automated gui interaction with native agents, 2025. URL https://arxiv.org/abs/2501.12326. ↗
  23. Yao Zhang, Zijian Ma, Yunpu Ma, Zhen Han, Yu Wu, and Volker Tresp. 2025b. Webpilot: A versatile and autonomous multi-agent system for web task execution with strategic exploration. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 23378–23386. find
  24. Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. 2023. Reflexion: Language agents with verbal reinforcement learning. Advances in Neural Information Processing Systems, 36:8634–8652.
  25. Paloma Sodhi, SRK Branavan, Yoav Artzi, and Ryan McDonald. 2023. Step: Stacked llm policies for web actions. arXiv preprint arXiv:2310.03720. — [arXiv:2310.03720](https://arxiv.org/abs/2310.03720) find
  26. Shuyan Zhou, Frank F Xu, Hao Zhu, Xuhui Zhou, Robert Lo, Abishek Sridhar, Xianyi Cheng, Tianyue Ou, Yonatan Bisk, Daniel Fried, and 1 others. 2023. Webarena: A realistic web environment for building autonomous agents. arXiv preprint arXiv:2307.13854.
  27. Yueqi Song, Frank Xu, Shuyan Zhou, and Graham Neubig. 2024. Beyond browsing: Api-based web agents. arXiv preprint arXiv:2410.16464. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yang-2026-webdart-dynamic-decomposition-re-planning/