HATS launches multi-agent AI system where six specialized AI assistants debate each other to improve decisions

HATS is a new software tool that runs meetings and planning sessions using six AI agents — each assigned a different thinking role (White Hat analyzes facts, Red Hat considers human impact, Black Hat identifies risks, Yellow Hat explores opportunities, Green Hat suggests alternatives, Blue Hat synthesizes conclusions) — that actively disagree with and challenge each other during discussions.

Unlike standard AI chatbots that give one confident answer, HATS forces structured conflict: agents debate trade-offs, surface blind spots, and present competing perspectives in real time. The system includes animated 3D avatars with synchronized speech, a Kanban task board that auto-dispatches work to agents, and integrations with tools like GitHub, Slack, Excel, and databases via the Model Context Protocol.

Teams can now stress-test startup ideas, plan products, or run retrospectives by watching AI agents argue through problems the way a diverse human team would — replacing async brainstorming sessions with live meetings where perspective clashes are the feature, not a bug. The tool supports OpenAI, Anthropic, and Google LLMs and runs on Node.js with no build step required.

HATS is open-source and self-hosted; users clone the repository, add their own LLM API keys (from OpenAI, Anthropic, or Google), and run `npm start` to access the web dashboard at localhost:3001. Setup scripts are provided for Windows, macOS, and Linux.