Overview

• Kiro transforms single prompts into detailed user stories with EARS-based acceptance criteria and sequences tasks for production-ready code.
• Event-driven hooks automatically trigger AI agents to generate tests, update documentation and optimize code whenever files change or on manual commands.
• The IDE leverages the Model Context Protocol to integrate Anthropic AI models in its preview and will add support for other providers later.
• AWS positions Kiro as a structured alternative to untracked ‘vibe coding’ practices to improve code quality, maintenance and traceability.
• Kiro is free during its preview period and will require a paid subscription afterward, with paying users’ content excluded from model training.