We use succinct data structures to represent our graphs. Succinct data structures are self indexing data structures that approach the information-theoretic mini

We use succinct data structures to represent our graphs. Succinct data structures are self indexing data structures that approach the information-theoretic minimum size of representation. Additionally, all the compute is on the local machine - we pass around highly compressed memory cores and do all of the orchestration (commit graph, permissions etc.) in the Hub. So far, the cost is very low so we are able to be generous in the free tier. For example, we have DBpedia available to clone on Hub now, we get about 10x compression on that so can serve it up really efficiently (clone to query in about 2 minutes with a normal enough connection). Hope is that we'll be able to keep it that way forever to facilitate public interest and open source work.