Show HN: Rebuno - An open-source runtime for production agents (github.com)

Rebuno has launched an open-source execution runtime designed specifically for production agents, offering a sophisticated framework that records every tool and Large Language Model (LLM) call as distinct, durable steps. This feature enables interrupted tasks to resume seamlessly from their last recorded point, enhancing efficiency and reliability in automated workflows. Users can manage step permissions—allowing, denying, or requiring human approval—adding a crucial layer of control in sensitive applications. This announcement is significant for the AI/ML community as it addresses key aspects of workflow management, including state durability and real-time event logging, which are vital for maintaining consistency in agent actions. Built on modern programming environments—requiring Go 1.26+, Python 3.11+, or Node 22+—Rebuno encapsulates advanced concepts such as policy management through YAML configurations, live streaming of incremental outputs, and comprehensive HTTP API support. The inclusion of a web UI for tracking executions and agent activities further bolsters its usability, making it an exciting tool for developers seeking to streamline AI deployments and enhance operational robustness.