Most of what Pipes did can be done locally, in a library. FeedReader for Windows did this.[1] Unfortunately, they recently announced
"Get ready for the brand new FeedReader! ... Check our site later this March for a new amazing web service, featuring industry-best RSS search and analysis.", which means it's becoming "cloud based" and probably has ads or worse. "rss.com" already went that way; they charge $6/month to read RSS feeds.[2]
Following RSS feeds on a continuing basis takes a lot of RSS polls. Most RSS feeds do not implement RSS in a way that allows getting only new items reliably.[3] The RSS "etag" mechanism is not reliable. Some sites with multiple servers and a load balancer have different etag values on each server. The "guid" field sometimes changes when the content hasn't changed. My experience is that nothing short of full text comparison eliminates duplicates properly. I wrote an RSS reader which does a MD5 of the text of each incoming message to throw out duplicates. Presumably the "pipes" system did something similar.
If RSS feed servers complied with the standard, there'd be less need for feed aggregation services.
[1] http://feedreader.com/
[2] https://www.rss.com/
[3] http://www.詹姆斯.com/blog/2006/08/rss-dup-detection