The main purpose of mongolog is to run logic programs with facts that are stored in a mongo database. Logic programs are translated into aggregation pipelines and executed server-side. This keeps the IO troughput small as the data is transformed and combined server-side.

The current implementation has some limitations and does not support the full ISO standard. One of the limitations is that rules cannot be asserted into the mongo database, but are baked into the aggregation pipeline. Another limitation is that recursion in rules is not allowed. However, many other aspects are supported by mongolog:

conjunctive and disjunctive queries;
negation as failure;
cut operator;
arithmetic expressions and operators;
data structures: atoms, numbers, compound terms, lists, variables;
comparison operators;
control and meta predicates;
findall and list predicates;
typechecking predicates;
database interaction (for facts only);
unification of compound terms (does not handle implicit instantiation yet).