While agent concepts from agent toolkits are used in mainstream software engineering, fully-fledged agent platforms remain niche. The conceptual gap between traditional object-oriented and service-oriented architectures and agent-based systems is key in making adoption challenging. In this article, we introduce Jadex V, a next-generation agent framework removing the need for explicit agent platforms, offering a lightweight, feature-oriented design enhancing configurability, integration, and usability. Jadex V provides a modular, feature-based infrastructure, simplifies Application Programming Interface (APIs), and unifies agent architectures under the active components paradigm. This approach allows seamless integration of Jadex as library or standalone framework, supporting a variety of agent types, like simple lambda agents, complex BDI agents, behavior trees and BPMN processes. Jadex V enhances performance and resource efficiency, making it suitable for different applications, including game AI, distributed processing and autonomous systems. Unlike traditional actor-oriented frameworks such as Akka or Orleans, which focus on message passing and concurrency, Jadex V bridges gaps between distributed infrastructure and intelligent agent architectures, enabling adaptive and autonomous multi-agent systems. This paper provides an in-depth exploration of Jadex V’s architecture, API, and feature set, illustrating its capabilities through concrete application examples. We discuss the advantages of using Jadex V for developing scalable multi-agent systems.