Right now we attempt to load everything at once - which caps out at 5,000 items and is really slow. We can do better by implementing pagination using count and offset.