Show HN: AgentRun – open-source, Self-improving Managed Agents

AgentRun is an open-source platform that enhances AI agents by adding self-improvement capabilities, durable execution, and a centralized control plane to reduce token usage and increase fault tolerance. The system learns from each run, storing useful information like notes and workflows to avoid redundant work, while maintaining run history even if interrupted and providing centralized monitoring for approvals, spending, and performance. Users can deploy it locally with a single command and start running agents across various environments like sandboxes, browsers, or mobile devices within minutes.