Cursor has expanded its long-running agents research preview to all Ultra, Teams, and Enterprise users, allowing these agents to autonomously tackle larger and more ambitious projects. This development follows Cursor’s research efforts, which revealed that using a custom harness significantly improves the performance and reliability of AI agents on complex tasks, from building a chat platform to refactoring systems. By planning tasks upfront and employing multi-agent checks, these agents have produced substantial pull requests efficiently and continue to refine their contribution to autonomous coding systems.
Cursor expands long-running agents research preview for Ultra, Teams, and Enterprise plans
