So, what you are saying is, that it doesn't exist? Why is no one building this? I understand, that there is less demand than for web search. But there still is a lot of demand from smart and skilled people, no?
But how does, e.g. Google Scholar estimate the number of citations? Do they only look at papers which are not behind a pay wall or do they buy a subscription to all journals?
It still buggles my mind, that the science loving google founders Larry and Brian haven't tackled this yet.