I'm impressed it seems to grabbed all of the images on a very JS heavy site with a carousel I just tried it on. Nice work.

One tiny UX thing to consider, instead of validating the string is a url, just try to load the site and if it fails, error then. I typed in a valid url without the "http://" part and the validation complained. You need to handle the "wait this isn't a valid site" case anyway, might as well kill two birds with one stone.