This article presents AgoraIA, an open-source framework designed to enable dynamic, multi-participant conversations by integrating multiple AI personas within a shared conversational environment. Unlike traditional single-agent LLM interactions, AgoraIA allows users to configure and deploy distinct AI agents that can engage in simultaneous, turn-based dialogues—both with human participants and with each other—through voice or text interfaces. The system supports a wide range of use cases, including formal panel discussions, interactive educational settings involving teachers and students, and informal conversations with friends or family members. By enabling diverse AI perspectives, roles, and communication styles, AgoraIA facilitates rich, context-aware exchanges that enhance engagement and collaboration. Key innovations include: (1) a modular architecture for creating customizable AI personas with unique roles, knowledge bases, and interaction styles; (2) support for synchronized turn-taking in mixed human–AI groups, ensuring structured yet fluid conversations; and (3) scalability across domains such as education, research, and decision-making. The effectiveness of AgoraIA is demonstrated through examples of its application in various professional and academic conversations involving users of different ages and experience levels.
Concha et al. (Thu,) studied this question.