It loads everything into memory, so it works well with <1GB datasets. The architecture could be changed to allow for larger datasets like yours, but that would

It loads everything into memory, so it works well with <1GB datasets. The architecture could be changed to allow for larger datasets like yours, but that would likely be a large undertaking and would probably be a paid feature.