What I like the most about Datasaur is the opportunity to design annotation guidelines that match the specifics of each individual project. As various text datasets may call for different interpretation standards, the system allows you to configure the annotation workspace according to these very standards. In this way, you reduce the risk of imposing similar classification principles to unrelated datasets. Moreover, I really like the fact that annotation instructions are always at hand near the actual records. Review collected by and hosted on G2.com.

Changing the annotation guidelines once the project is already started may be quite difficult. With changing requirements, it could be useful to have some clear tools to indicate which instructions were modified and inform all participants about the changes. Review collected by and hosted on G2.com.