An open research foundation dedicated to pushing the boundaries of computer use agents. We bring together researchers and developers to ensure this technology benefits humanity.
Our research spans the fundamental challenges of building intelligent agents that interact with digital environments.
How do agents perceive and interpret graphical interfaces? We study screen comprehension, UI element recognition, and spatial reasoning.
Research into how agents decompose complex tasks, plan action sequences, and adapt strategies when encountering unexpected states.
Developing rigorous methodologies to measure agent capabilities, compare approaches, and track progress toward general computer use.
Ensuring agents act within intended boundaries. Research on oversight mechanisms, failure modes, and alignment with human values.
Understanding how agents maintain context, recover from errors, and complete workflows that span extended time periods.
Studying effective interaction patterns between humans and agents. When should agents act autonomously vs. seek guidance?
Computer use agents represent a fundamental shift in how humans interact with technology. We believe this transition must be guided thoughtfully.
Conduct and publish foundational research that pushes the boundaries of what computer use agents can achieve. Open problems require open collaboration.
Develop open-source tools, datasets, and benchmarks that accelerate progress across the entire field. Knowledge should be shared, not siloed.
Create spaces for researchers, engineers, and policymakers to discuss the most important questions about intelligence and its role in society.
Guide the development of computer use agents so their benefits are widely distributed. Technology this powerful should serve humanity, not just a few.
Facilitate rigorous discussion on safety, ethics, and governance. The decisions we make now will define how this technology evolves.
Source & Research
Research Community
Organization
Governance
We bring together diverse perspectives to tackle the most important questions about intelligence and computing.
Collaborate on open problems, publish with us, and access shared infrastructure. We support both independent researchers and academic institutions.
Contribute to open-source projects, help build evaluation frameworks, and turn research insights into practical tools for the community.
Engage with technical experts to understand capabilities and risks. Informed policy requires deep collaboration between technologists and governance experts.
Partner with us to explore responsible deployment, fund critical research, and help shape standards for the emerging field of computer use agents.
Join researchers and developers working to ensure intelligent agents benefit all of humanity.
Get in Touch