I built a tool that forces 5 AI to debate and cross-check facts before answering (github.com)

KEA Research has introduced a groundbreaking multi-AI collaboration platform that facilitates fact-checking and verification by orchestrating responses from up to five different AI models. This tool tackles the prevalent issue of inconsistent answers, wherein each AI provides responses with equal certainty, potentially leading to misinformation. By employing a structured four-step pipeline—initial response gathering, peer refinement, evaluation of consensus, and final synthesis—KEA ensures that only verified, consensus-backed information is presented to users, significantly enhancing the reliability of AI-generated answers. This innovation is vital for the AI/ML community as it addresses the challenge of information accuracy in a landscape where users are increasingly reliant on AI for decision-making. The platform supports major AI providers like OpenAI, Anthropic, and Google, allowing diverse perspectives while flagging disputed claims. With capabilities to analyze images, export findings in multiple formats, and support 75 languages, KEA Research extends its utility across various domains, from academic research to business strategy analysis. The project's name, inspired by the intelligent New Zealand parrot, underscores a commitment to collaborative problem-solving and advancing the field of AI through transparency and consensus.