We don't have any hard-and-fast throttles in place for the PGCR service, but we ask that you try to limit it to under 20 requests/second. Indeed, I've heard of some sites that increment the PGCR URL. It's a bit of an extreme undertaking, but the benefit you get is total informational awareness. You should hit up @Steffwiz, he is a developer on one of the sites that does this en masse scraping, and perhaps you could collaborate on an effort together - and save yourself a lot of bandwidth costs and trouble if you guys can work together with that already scraped dataset! What kind of use case were you thinking of? Maybe it's something that can be done without having to scrape all PGCRs? If not, definitely hit up someone like Steffwiz before bothering to re-grab all of this data if you can... it could be that your use cases are similar enough that you can collaborate somehow!