Elamur Research⌕
☰
home / papers / 2026 / Demonstrating ViviDoc: Generating Interactive Documents through Human-Agent Collaboration

Demonstrating ViviDoc: Generating Interactive Documents through Human-Agent Collaboration

Paper · 2026 · Yinghao Tang, Yupeng Xie, Yingchaojie Feng et al.

Summary

Interactive explanatory articles, with sliders and visuals that readers can manipulate, help people learn, but they are expensive to build because they need both subject expertise and web programming. Simply asking an LLM agent to write such an article gives output that is hard to steer or verify. ViviDoc is a demo system that turns a single topic into an interactive educational HTML document using three agents:

  • a Planner, which drafts a plan;
  • an Executor, which writes each section's text and then its HTML/CSS/JavaScript visualization, retrying on failure;
  • an Evaluator, which checks coherence and HTML validity.

The plan uses an editable intermediate format called DocSpec. It splits the document into knowledge units, and each visualization is described through State, Render, Transition and Constraint (SRTC), an approach grounded in visualization theory. Educators can revise the plan directly or through a chat assistant before any code is generated. The authors also assembled a benchmark of 101 real interactive documents. In blind expert ratings, ViviDoc clearly beat single-call generation with the same model. The work shows how structured specifications can make LLM generation controllable.

Key points

  • Benchmark: 101 real-world interactive documents collected from more than 60 websites, spanning 11 domains.
  • Expert blind evaluation covered 10 sampled topics and 20 documents, rated by 3 experts on a 1–5 scale. Both systems used Gemini 3.0 Flash, and human review was disabled for the comparison. ViviDoc outscored the naive agent on every dimension:
    • Content Richness: 4.17 vs 2.07
    • Interaction Quality: 4.00 vs 2.40
    • Visual Quality: 3.73 vs 2.37
  • In SRTC, State holds variables and their types, Render maps state to visual elements, and Transition maps user actions to state changes. Constraint encodes the pedagogical invariant, for example that C/D stays near π whatever the radius.
  • The user study was small: 3 participants did tasks in math, algorithms and physics. They gave ease of learning and ease of use 5.0 each, and rated DocSpec alignment with their expectations 4.67.
  • Humans can intervene twice: after planning, by editing the DocSpec, and after generation, by requesting revisions.

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

🎓National University of SingaporeSingapore · Singapore180 works · 34,304 citations🎓Hong Kong University of Science and TechnologyHong Kong · China142 works · 25,136 citations🎓Zhejiang UniversityHangzhou · China172 works · 12,988 citations🎓University of VirginiaCharlottesville · United States36 works · 9,730 citations🎓The Hong Kong University of Science and Technology (Guangzhou)Guangzhou · China53 works · 1,367 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksComputer Vision and Pattern Recognition · 2,869 worksInformation Systems · 1,437 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 works

Cite

Cite as Yinghao Tang, Yupeng Xie, Yingchaojie Feng, Tingfeng Lan and Wei Chen. 2026. Demonstrating ViviDoc: Generating Interactive Documents through Human-Agent Collaboration. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-demo.79

@inproceedings{tang2026demonstrating,
  title = {{Demonstrating ViviDoc: Generating Interactive Documents through Human-Agent Collaboration}},
  author = {Yinghao Tang, Yupeng Xie, Yingchaojie Feng, Tingfeng Lan and Wei Chen},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-demo.79},
  url = {https://doi.org/10.18653/v1/2026.acl-demo.79}
}

References

24 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Yutong Bian, Xianhao Lin, Yupeng Xie, Tianyang Liu, Mingchen Zhuge, Siyuan Lu, Haoming Tang, Jinlin Wang, Jiayi Zhang, Jiaqi Chen, and 1 others. 2025. You don’t know until you click: Automated gui testing for production-ready software evaluation. arXiv preprint arXiv:2508.14104. — [arXiv:2508.14104](https://arxiv.org/abs/2508.14104) find
  2. Donald Bitzer, Peter Braunfeld, and Wayne Lichtenberger. 2007. Plato: An automatic teaching device. IRE Transactions on Education, 4(4):157–161. find
  3. John Branch. 2012. Snow fall: The avalanche at tunnel creek. The New York Times. https://www.nytimes.com/projects/2012/ snow-fall/index.html. from documents. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 20552– 20570. ↗
  4. Kanika Goswami, Puneet Mathur, Ryan Rossi, and Franck Dernoncourt. 2025. Plotgen: multi-agent llm-based scientific data visualization via multimodal retrieval feedback. In Companion Proceedings of the ACM on Web Conference 2025, pages 1672–1676. find
  5. Fred Hohman, Matthew Conlen, Jeffrey Heer, and Duen Horng Chau. 2020. Communicating with interactive articles. Distill. find
  6. Boyan Li, Yiran Peng, Yupeng Xie, Sirong Lu, Yizhang Zhu, Xing Mu, Xinyu Liu, and Yuyu Luo. 2026. Deepeye: A steerable self-driving data agent system. arXiv preprint arXiv:2603.28889. — [arXiv:2603.28889](https://arxiv.org/abs/2603.28889) find
  7. Xin Liang, Xiang Zhang, Yiwei Xu, Siqi Sun, and Chenyu You. 2025. Slidegen: Collaborative multimodal agents for scientific slide generation. arXiv preprint arXiv:2512.04529. — [arXiv:2512.04529](https://arxiv.org/abs/2512.04529) find
  8. Yi-Cheng Lin, Kang-Chieh Chen, Zhe-Yan Li, Tzu-Heng Wu, Tzu-Hsuan Wu, Kuan-Yu Chen, Hung-yi Lee, and Yun-Nung Chen. 2025. Creativity in llm-based multi-agent systems: A survey. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 27572–27595. find
  9. Tamara Munzner. 2014. Visualization Analysis and Design. A K Peters Visualization Series. A K Peters/CRC Press. find
  10. Yanzhe Chen, Kevin Qinghong Lin, and Mike Zheng Shou. 2025a. Code2video: A code-centric paradigm for educational video generation. arXiv preprint arXiv:2510.01174. — [arXiv:2510.01174](https://arxiv.org/abs/2510.01174) find
  11. Theodor Holm Nelson. 1965. Complex information processing: a file structure for the complex, the changing and the indeterminate. In Proceedings of the 1965 20th national conference, pages 84–100. find
  12. Yiyu Chen, Yifan Wu, Shuyu Shen, Yupeng Xie, Leixian Shen, Hui Xiong, and Yuyu Luo. 2025b. Chartmark: A structured grammar for chart annotation. In 2025 IEEE Visualization and Visual Analytics (VIS), pages 311–315. IEEE. find
  13. Bo Pan, Yixiao Fu, Ke Wang, Junyu Lu, Lunke Pan, Ziyang Qian, Yuhan Chen, Guoliang Wang, Yitao Zhou, Li Zheng, and 1 others. 2025. Vis-shepherd: Constructing critic for llm-based data visualization generation. arXiv preprint arXiv:2506.13326. — [arXiv:2506.13326](https://arxiv.org/abs/2506.13326) find
  14. Michelene T. H. Chi and Ruth Wylie. 2014. The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational Psychologist, 49(4):219–243. find
  15. Eric Roston and Blacki Migliozzi. 2015. What’s really warming the world. Bloomberg Business. https://www.bloomberg.com/graphics/ 2015-whats-warming-the-world/. ↗
  16. Victor Dibia. 2023. Lida: A tool for automatic generation of grammar-agnostic visualizations and infographics using large language models. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations), pages 113–126. find
  17. Yinghao Tang, Xueding Liu, Boyuan Zhang, Tingfeng Lan, Yupeng Xie, Jiale Lao, Yiyao Wang, Haoxuan Li, Tingting Gao, Bo Pan, and 1 others. 2026. Igenbench: Benchmarking the reliability of text-to-infographic generation. arXiv preprint arXiv:2601.04498.
  18. Douglas C Engelbart. 2023. Augmenting human intellect: A conceptual framework. In Augmented education in the global age, pages 13–29. Routledge. find
  19. Bret Victor. 2011. Explorable explanations. http://worrydream.com/ExplorableExplanations/. Accessed: 2026-02-27. ↗
  20. Akash Ghosh, Aparna Garimella, Pritika Ramu, Sambaran Bandyopadhyay, and Sriparna Saha. 2025. Infogen: Generating complex statistical infographics Martin Wattenberg, Fernanda Viégas, and Moritz Hardt. 2016. Attacking discrimination with smarter machine learning. Google Research. https://research.goo… ↗
  21. Yupeng Xie, Yuyu Luo, Guoliang Li, and Nan Tang. 2024. Haichart: Human and ai paired visualization system. arXiv preprint arXiv:2406.11033. — [arXiv:2406.11033](https://arxiv.org/abs/2406.11033) find
  22. Yupeng Xie, Chen Ma, Zhenyang Wang, Liangwei Wang, Jiayi Zhu, Chuxuan Zeng, Zhouan Shen, Boyan Li, and Yuyu Luo. 2026a. Datamagic: Transforming tabular data into data insight video. arXiv preprint. find
  23. Yupeng Xie, Zhenyang Wang, Liangwei Wang, Jiayi Zhu, Zhouan Shen, and Yuyu Luo. 2026b. Datamagic: Authoring data videos through declarative multi-agent orchestration. arXiv preprint. find
  24. Yupeng Xie, Zhiyang Zhang, Yifan Wu, Sirong Lu, Jiayi Zhang, Zhaoyang Yu, Jinlin Wang, Sirui Hong, Bang Liu, Chenglin Wu, and 1 others. 2025. Visjudge-bench: Aesthetics and quality assessment of visualizations. arXiv preprint arXiv:2510.22373. \#\# Ethical Considerations — [arXiv:2510.22373](https://arxiv.org/abs/2510.22373) find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/tang-2026-demonstrating-vividoc-generating-interactive-documents/