FRAME (A Model-Based Foundry for Engineering, Adapting, and Assuring the Quality of AI Agents) is a European research consortium project focused on building tools and methods to engineer, deploy, and assure the quality of AI agent systems.

The consortium includes Gradiant (coordinator), Process Talks, CUNI, PAL Robotics, SAS/FISEVI, SIMULA, DCU, BU, UvA, and F6S, each contributing to different work packages spanning requirements, DSL/meta-models, integration, and pilot validation.

The project is structured around three pilot use cases: robotics (PAL Robotics), business process automation (Process Talks/PTAP), and healthcare (SAS/FISEVI).

Core project values include AI Act compliance, data sovereignty, and open-source, self-hostable, Kubernetes-deployable tooling.

IRATEN Solutions (IRS) leads Work Package 6 (WP6), which focuses on integrating, deploying, operating, and assuring components built by other work packages, including AgentOps capabilities, rather than developing new agent systems.

Within WP6, IRS is defining framework requirements and specifications (T6.1), gathering NFRs and use-case feedback from the three pilots, and selecting observability/traceability tooling for T6.3, notably recommending Langfuse with OpenLLMetry/Traceloop as a complementary layer.

IRS also contributes to T6.4’s agentic copilot interface work. Beyond WP6, IRS leads Task 8.3 (Standardisation & Policy Contribution) under WP8, engaging with standards bodies such as ISO/IEC SC 42, CEN-CENELEC JTC 21, ETSI, IEEE, GAIA-X, and BDVA/DAIRO, and producing policy briefs, EC contributions, and workshops across the project timeline.