There exist many distributed storage systems that are mature and reliable. These systems are not fully adapted to the needs of an open astroparticle community mainly because they do not follow the Open Archival Information Systems (OAIS) standard. Moreover they require adaptations by each of the data centres at which they are installed. We introduce GAMAS, a novel distributed OAIS that tackles the problem of different technologies used at the various DCs. Instead of imposing the requirements of GAMAS to the DCs, we allows them to simply provide a python interface (or plugin) to their storage. This allows GAMAS to be easily deployed on top of different architecture and technologies, and to transparently allow users to retrieve data for processing, wherever they are. A metadata browsing system is incorporated within GAMAS and allows DCs as well as anonymous users to retrieve datasets based on high-level queries. GAMAS’ central database can be reconstructed from the archived data, which makes the system robust against corruption. We will expose the current status, architecture and functionalities of GAMAS and also detail the current test case which stores about 0.6 PB of data from the FACT experiment at separate data centres.