In architectural education and practice, students, teachers and architects increasingly rely on online repositories with architectural information. This includes product model data, exemplary architectural projects and technical documentation, in a wide variety of formats. Unfortunately, the major part of this architectural content exists in individual repositories and they are not accessible to geographically distributed professionals and students who require them. Some of them are freely accessible, while others require registration or even subscription. Since future architects and engineers will increasingly rely on product model repositories as their project database, much work is to be done to make such repositories easy to use, intuitive and effective.The last decade has also shown an increased application of 3D models. On the one hand, design projects rely more and more on 3D modeling techniques and on the other hand, a growing amount of information is shared using 3D content, such as product and building models. To this extent, a variety of software applications is applied, ranging from generic 3D and CAD software, over dedicated architectural design applications. There are already several 3D Model repositories available on internet. Some repositories allow for a structured search, e.g. looking for style, function or location, while many others are generic in nature, providing only a very crude classification, when considering architecture. They might have a “building” category, but seldom use additional classification. This severely limits their usefulness in an architectural context. While many architectural practices usually avoid sharing their models, to respect confidentiality and privacy of their clients, designers and students are more eager to make their models available. However, to be able to accurately retrieve models – apart from endlessly browsing through the repositories – metadata are required. This article describes a suggested approach to properly classify and structure metadata for Architectural 3D Content. To facilitate this, the IEEE Learning Object Metadata (LOM) standard is applied. To clarify the specificities of 3D content, a general description will be given on 3D modeling and CAD and more particular about the approach with Building Information Modeling and its impact on data exchange, through the development of the Industry Foundation Classes. We will also present a summarized overview of available 3D content repositories and 3D file formats, as they are used to derive the common characteristics of 3D content. This will give the main outline of the proposed metadata scheme. This scheme is confronted and augmented with the structure derived from IFC, to ensure applicability with BIM applications. Finally, the proposed metadata scheme is presented, following the LOM standard and its potential usage is discussed. To motivate the chosen series of metadata fields, different approaches are followed. On the one hand a metadata scheme is developed in the course of the MACE project and in this scheme the Media taxonomy is set up. To assess the required metadata fields, a group of common fields is derived from a series of categorized existing repositories. This collection is used to derive already applied metadata fields. In addition, a collection of file formats for 3D and CAD models is described, as to derive common characteristics of 3D models. This article focuses on the extension of the Domain Taxonomy.