This chapter is a response to the increasing demand for a versatile AI dataset repository in a rapidly evolving landscape that necessitates frugal and robust AI solutions for defence applications. In an era characterized by an exponential growth in data generation, the need for a repository that can accommodate diverse and extensive datasets has become paramount. The evolving nature of defence and security challenges, marked by the reliance on AI for critical decision-making, further underscores the urgency of this repository. With a wealth of non-cooperative and cooperative tracking data, electro-optical data, radio signals, and more at play, the chapter addresses the imperative to provide a secure, accessible, and scalable repository that aligns datasets with specific user requirements and use cases. This comprehensive repository design acknowledges not only the massive volumes of data at hand but also the importance of long-term data preservation, access control, and data integrity. It offers a methodological blueprint for building a robust AI dataset repository capable of meeting the ever-increasing demands of a dynamic AI landscape. In this context, the chapter includes the methodological approach for the creation of a specialized AI dataset repository, which is being developed within the framework of the EU-funded FaRADAI project (GA no. 101103386) and aims at advancing AI technology for defence applications. The FaRADAI Dataset Repository (FDR) plays a pivotal role in facilitating collaboration among project partners by serving as a centralized hub for secure and efficient data storage and access.