I have an already built a scraping service which is very dynamic and been working for 18 months, so its very stable and proven. I can configure and create new scrapers through an interface I've built which control all aspects of scraping (how often, from what sources, what fields to scrap and more).