Currently, on a global scale, there is an endeavor to create new ways of sharing knowledge and promoting learning opportunities mostly based on educational content published on the Web. Moreover, the COVID-19 pandemic has made visible efforts oriented to support online learning. In such a context, Open Educational Resources (OER) provide a greater opportunity for all people to access digital materials for teaching and learning, without regarding their geographical locations or economic limitations. However, there are some concerns about OER search and visibility. Indeed, OER search experience for non-expert users demands time and effort to find educational content that matches their specific requirements for subject context, educational level, format, pedagogical approach, associated license, among other characteristics. To address this problem, structured metadata can be embedded in web pages using annotations in an encoding format and supporting vocabulary. Schema.org introduced a standardized vocabulary for providing a large range of terms for describing any web content. The most important search engines have adopted these vocabulary terms to enrich the search results. Nevertheless, using the vocabulary to ensure OER web pages visibility requires knowledge about the vocabulary and the JSON-LD format. To overcome this barrier, we developed a web application whose functionality allows OER producers to generate the JSON-LD structure that should be inserted in the head of the educational web page. In this paper, we present the development details of this JSON-LD generator for automatic OER annotation, whose foremost goal is to enable users to choose among the most relevant educational descriptors for educational resources in different categories as a web page, course, article, book, video, and image. Our in-progress research takes one more step towards a whole OER environment that improves affordability, accessibility, and usability.
