Social media has become important for signal processing research due to the interesting content consisting of status updates and relations which give clues about social ties. Thus, it is important to lower the barrier to access this kind of data for researchers of natural language processing, social graph analysis and machine learning. However, accessing this data can usually be tricky and may pose difficulties. This work describes a distributed system for collecting, storing and serving social media data using the computing resources shared by its own volunteer users. We used the system extensively and saw promising results in terms of time and space efficiency.