Abstract
Current Agent Toolkits are too restricting, by preventing their Users to choose the Agent?s Decision Making Process and Testing/Deployment Environment. I propose a flexible Framework capable of integrating Embodied Intelligent Agents with any Virtual Environment. By maintaining a clear and abstract structure of the interaction between the Decision Making Process Centre and the Environment, Framework Users will also be able to choose each Agent?s Decision Making Process. The Agent will always maintain its full feature specification, not mattering what Environment it is. This article describes the proposed Framework?s basic concepts, architecture and several implementation approaches, and presents Use Cases that will prove the soundness of the Framework and its capabilities.