Cursor, a leading AI-driven code editor, has significantly enhanced its capabilities, enabling developers to run multiple AI agents concurrently across diverse codebases. The update introduces asynchronous subagents, improved worktrees, and multi-root workspaces, streamlining complex development workflows. Prominent developer Lee Robinson expressed enthusiasm for the new features, stating: > "Extremely excited about this Cursor ship. It's now way easier to run many agents at the same time (with async subagents and improved worktrees). And you can run agents across many repos (e.g. frontend and backend!)."
The core of this advancement lies in the introduction of asynchronous subagents, which allow Cursor to parallelize requests rather than queuing them. Through the /multitask command, the system can break down larger development tasks into smaller chunks, assigning them to a fleet of subagents that operate simultaneously. This significantly boosts efficiency by enabling parallel execution of various coding assignments.
Further enhancing developer productivity are the improved worktrees, now integrated within the Agents Window, and multi-root workspaces. Worktrees facilitate isolated task execution across different branches, leveraging Git worktrees to prevent conflicts when multiple agents are active. Concurrently, multi-root workspaces enable a single agent session to target multiple project folders, allowing seamless cross-repository changes spanning frontend, backend, and shared libraries without constant agent retargeting.
These features are part of the broader Cursor 3.0 release, which aims to provide a unified workspace for building software with AI agents. Cursor 3.0 represents a strategic shift towards an agent-centric interface, pushing the boundaries of AI-assisted development towards more autonomous and collaborative agent fleets. The company continues to invest in transforming the coding experience, preparing for a future where AI agents play an even more integral role in software creation.
