Dagu is a self-contained workflow orchestration engine similar to Airflow or Cron, designed for small teams requiring no database. It allows defining DAGs in declarative YAML format with a visual Web UI for monitoring and retrying pipelines. The engine natively supports AI agents through Model Context Protocol integration, enabling AI agents to manage and control workflows. It includes built-in agent harnesses and actions for constructing AI-driven workflows alongside traditional shell, Docker, and Kubernetes tasks.