I've looked into this problem briefly myself, and as a PhD student who runs an academically oriented start-up/web app, it's something I'd be very interested in

I've looked into this problem briefly myself, and as a PhD student who runs an academically oriented start-up/web app, it's something I'd be very interested in seeing happen. In principle it should be quite easy to get working, at least for a few select journals (in my case it's the arXiv I'm interest in). However, the one stumbling block that I see is an easy and effective way to extract citations. An important filter for whether a paper is relevant to me would be whether it cites some seminal papers in my field. There are services out there which index citations (SPIRES in my field[1]) but there is always a bit of a delay before they are updated. Using them would ruin the real-time nature of the service. The only way I see around that is to extract the citations from the papers in the journal feed as they come in. But now the service has gone from a weekend project to a large undertaking that would require a substantial infrastructure to work. I'll see how this thread continues to evolve, but I'd like to have further discussions about making this a reality. [1] http://www.slac.stanford.edu/spires/ http://www.slac.stanford.edu/spires/