As of 2012, ArgDB as a corpus is simply a subset of resources available on the Argument Web and accessible via the AIFdb interface.
ArgDB has grown out of AraucariaDB, providing online access to a corpus of analysed arguments. All those arguments that were available in AraucariaDB have been transcoded and are now available in ArgDB. (AraucariaDB is still maintained, however, as it is directly accessible from the Araucaria software). As a part of the Dialectical Argumentation Machines project, this resource will be expanded. But perhaps the most significant difference is that ArgDB is built on AIF representations and WWAW infrastructure.
ArgDB also provides a REST-ful API, a description of which is available. This page also describes how to deploy your own ArgDB instance.