Right now you seem to only be able to pull website URLs via AI, costing AI credits, but this is something that can be done with a crawler like screaming frog with no AI involvement - leading to the same output, but vastly cheaper.