CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents
Xiangru Jian 2 , * Shravan Nayak 1 , 3 , 4 , * Kevin Qinghong Lin 6 Aarash Feizi 3 , 5 Kaixin Li 7 Patrice Bechard 1 Spandana Gella 1 , 5 Sai Rajeswar 1 , 3 1 ServiceNow 2 University of Waterloo 3 Mila 4 Université de Montréal 5 McGill University 6 University of Oxford 7 National University of Singapore * Equal Contribution · 2026
Read CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents by Xiangru Jian 2 , * Shravan Nayak 1 , 3 , 4 , * Kevin Qinghong Lin 6 Aarash Feizi 3 , 5 Kaixin Li 7 Patrice Bechard 1 Spandana Gella 1 , 5 Sai Rajeswar 1 , 3 1 ServiceNow 2 University of Waterloo 3 Mila 4 Université de Montréal 5 McGill University 6 University of Oxford 7 National University of Singapore * Equal Contribution (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-24440.txt.
Introduction
Jian and co-authors introduce CUA-Suite, a large resource of human demonstration videos for computer-use agents. Progress on desktop agents is held back by a lack of continuous, high-quality recordings; sparse screenshots miss how actions unfold over time. At its core, VideoCUA holds about 10,000 expert tasks across 87 applications, with 30 fps screen video, cursor traces, and layered reasoning notes. That is roughly 55 hours and six million frames. The suite also adds UI-Vision and other resources to support agent training and testing.
Catalog record
| CREATOR | Xiangru Jian 2 , * Shravan Nayak 1 , 3 , 4 , * Kevin Qinghong Lin 6 Aarash Feizi 3 , 5 Kaixin Li 7 Patrice Bechard 1 Spandana Gella 1 , 5 Sai Rajeswar 1 , 3 1 ServiceNow 2 University of Waterloo 3 Mila 4 Université de Montréal 5 McGill University 6 University of Oxford 7 National University of Singapore * Equal Contribution |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 185275 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.24440