The Eclipse Foundation has unveiled the Agent Definition Language (ADL) as part of the Eclipse LMOS platform, marking a significant milestone in enterprise AI agent development. This open-source initiative stems from the necessity at Deutsche Telekom to deploy artificial intelligence across multiple European territories while leveraging their existing Java infrastructure.
ADL is designed to empower business users by allowing them to define AI agent behavior without the need for complex coding. It serves as a structured, model-agnostic framework that simplifies the intricacies of traditional prompt engineering, bridging the gap between technical and non-technical stakeholders in AI development.
The Eclipse LMOS platform, which now integrates ADL, provides a comprehensive environment for developing and orchestrating AI agents. Running on Kubernetes with a JVM-native agent framework and a Kotlin runtime, the platform includes an orchestration layer for managing the lifecycle of AI agents. By aligning with enterprise technologies like Kubernetes and Istio, Eclipse LMOS challenges the current Python-dominated landscape of AI tooling.
One of the key features of ADL is its ability to separate business logic from prompts, enabling a continuous iteration of agent rules while maintaining robust version control. This feature provides a substantial advantage for enterprises aiming to build scalable, multi-agent systems without relying on proprietary solutions.
Already in production, the integration of ADL within the LMOS platform has been tested thoroughly at Deutsche Telekom, making it one of the largest enterprise AI agent deployments in Europe. The platform has managed millions of interactions, proving its capability and reliability in handling large-scale enterprise applications.

Collections

The Eclipse Foundation launched the Agent Definition Language (ADL) as part of the Eclipse LMOS platform, enabling business users to define AI agent behavior without complex coding. Built on Java and Kubernetes infrastructure, the platform provides a model-agnostic framework that separates business logic from prompts and supports version control. Already deployed at Deutsche Telekom, it has handled millions of interactions across European territories, representing one of the largest enterprise AI agent deployments in Europe.

Eclipse Foundation Introduces Agent Definition Language for Enterprise AI Development