I know next to nothing about node.js; I managed to install it, as well as npm and an app called forever that will run river.js in the background. (Thanks to Chris Dadswell and Andrew Shell for their blog posts about this.)
If you know enough to install apps in Linux, grab the above, clone river 4 via git, and run
node river.js from within the river4 folder. If everything works, you can run
forever start river.js in the background.
Honestly, the hardest part was getting my Amazon S3 bucket to work. I learned via Stack Overflow that in order to map river.jeffreykishner.com via CNAME to my bucket, I had to name the bucket river.jeffreykishner.com. That was the only major hurdle.
I also learned that I needed to store the three environmental variables in .bashrc so that they would stay permanent.
The cool thing about River 4 is that one can use include to reference an OPML file that is not stored in your S3 bucket. As an experiment, I uploaded an OPML file that includes a link to an OPML file hosted in Fargo. Sure enough, the feeds in the OPML file in Fargo appear in my river, so in the future, all I have to do is add a new feed to my list in Fargo instead of having to re-upload an OPML file to Amazon everytime I want to add a new feed to a tab.