As more and more applications are data-driven, being able to share semi-structured data seamlessly becomes an absolute necessity. Building social networking applications or semantic wikis, for instance, relies today mostly on the capability to share, integrate and redistribute heterogeneous pieces of data across large sets of users. Supporting semi-structured data sharing is especially crucial in a learning context, where students are often expected to contribute highly-structured content (e.g., scientific reports, quizz answers) that is then compared, merged or integrated. However, most of today’s open data sharing infrastructures strongly rely on the availability of third-party online resources (such as DBpedia http://dbpedia.org/) that are not always available in rural or remote contexts. Our project will help teachers and students in two specific ways: i) we will build an open software infrastructure to more easily share and integrate semi-structured data locally in ad-hoc contexts and ii) we will build mechanisms to then seamlessly integrate the contributed data into global data sharing platforms.