Such a system might be constructed for one’s own scraping needs by taking any one of the current frontend/backends and customizing that behavior such that it could mitigate issues or ingest/ignore data based on your own inputs as well; such that your model could be “riding along on a human surfboard with human guidance”