Show HN: God – real-time control room for LLM agent societies (EMNLP 2026 demo) (github.com)

A new AI demonstration tool called GOD has been accepted for the EMNLP 2026 conference, revolutionizing the way language-agent societies can be interacted with. Unlike traditional simulations that offer a passive viewing experience, GOD allows users to take an active role by pausing time, altering agent interactions, and exporting scenarios for others to replicate. Key features include a no-code setup wizard for easy configuration of models, scenarios, and agents directly from a browser, making it accessible to a broad audience, including those without programming experience. This tool is significant for the AI/ML community as it bridges the gap between simulation and real-time control, enabling researchers to conduct more dynamic experiments. It supports custom environments, allowing users to create detailed societal setups equipped with varying weather conditions, events, and resident profiles that can significantly affect agent behavior. With capabilities such as real-time intervention, localized data management, and integration with OpenAI-compatible models, GOD has the potential to facilitate new research paradigms while expanding the boundaries of agent-based modeling in AI.