Cloud Agents with Computer Use

Cursor has launched a new capability for its cloud agents: the ability to use software they create to test changes and demonstrate their work. This represents a significant expansion in autonomous development capabilities.

How It Works

After onboarding onto your codebase, each cloud agent runs in its own isolated VM with a complete development environment. This sandboxed approach ensures safety and consistency while allowing agents to:

• Execute and test code changes in real time
• Capture video demonstrations of their work
• Generate screenshots and detailed logs
• Validate functionality before proposing changes

Deliverables and Review Process

Cloud agents now produce merge-ready pull requests accompanied by rich artifacts that streamline code review:

• Videos showing agent actions and testing procedures
• Screenshots capturing key moments and results
• Comprehensive logs for debugging and verification

This makes it possible to quickly understand and review agent-generated changes without manually running or testing the code yourself.

Availability

The feature is available across Cursor's entire platform ecosystem:

• Web interface
• Desktop application
• Mobile clients
• Slack integration
• GitHub integration

Developers can get started by visiting cursor.com/onboard to watch the agent configure itself and record a demo, or read the full announcement.