Show HN: Quoracle: Self-replicating multi-LLM-consensus agents (Elixir) (github.com)

Quoracle, a new Phoenix LiveView application, has been introduced to the AI/ML community, allowing users to create hierarchical systems of agents that operate by consensus across multiple language models (LLMs). This innovation enables decision-making based on agreement among various models, enhancing the reliability of outputs compared to reliance on a single LLM. The platform supports the creation of agents that can spawn child agents, communicate, and persist their states in PostgreSQL, all accessible via a real-time browser dashboard. The significance of Quoracle lies in its potential to advance multi-agent orchestration and consensus-driven AI applications. By harnessing insights from multiple models, users can tackle complex tasks more effectively, leveraging diverse perspectives for more accurate results. The technical framework requires a variety of components such as API keys for different LLM providers and PostgreSQL for data management. It also incorporates features like role-specific task assignments, customizable cognitive styles, and global constraints to streamline agent operations. Quoracle positions itself as a tool for researchers and developers eager to delve into the intricacies of multi-agent systems, moving beyond simple chat interfaces to complex, consensus-based task execution.