Collections of cultural heritage documents are most of the time digitized images. These precious documents are now available on the internet and need manual annotations to make their content accessible and exploitable. Manual annotations are expensive and tedious; hence the use of a collaborative annotation space and the reuse of users’ experiences, by tracing their actions -during the annotation process- in order to offer an assistance based on these traces is necessary. In this article we present a digital archive model and prototype of a collaborative system to annotate online ancient manuscripts. The application offers an assistant for semi-automatic annotation, and a tracing system that saves traces of important actions in order to reuse them in a recommender system afterwards.