Thank you! and correct, we have long running agents that Source → Crawl → Evaluate → Index → Retrieve → Contract this data. O(1 crawler) for the web vs O(field

Thank you! and correct, we have long running agents that Source → Crawl → Evaluate → Index → Retrieve → Contract this data. O(1 crawler) for the web vs O(fields × jurisdictions) for real-world data. So we have to build a lot of unique crawlers.