Annotation is the process of supplementing data with additional information that was not part of the actual observation, but reflects post-facto comments and associations made by a user who analyzes the data. While annotation management systems are emerging in the field of relational data, such systems for scientific applications, where there is a wide heterogeneity in the types of annotable data, are almost nonexistent. In this demonstration paper, we describe Graphitti, a tool that (i) allows a user to annotate a wide variety of scientific data, and (ii) allows a user to query data and their annotations in a seamless manner.