It is a static site feed aggregator primarily designed to go with GitHub Pages. I host one to aggregate my own writing on different sites. I think it may fit your use case because your Osmosfeed site itself outputs a single Atom feed. So, for example, if I have an Osmosfeed site that aggregates feeds 1, 2, and 3, the Osmosfeed site has a single feed which will include the three individual feeds. Mine has about 10-12 feeds and it has worked perfectly thus far with no issues. Not sure if it would have problems at higher numbers.
Do you have a link to YOShInOn (is it open source/are there more details somewhere)? I'm curious
I do have to make a slide deck or web page about it, I think there’s a lot of information that would be useful to other people who might want to try something similar but I have to make up my mind about directions for it to go.
On one hand I like it as an RSS reader and there is a direction where it gets mashed up with a bookmarks manager and is a sort of “consumer” product with the serious caveat that a self-hosted version is going to require you to classify maybe 200 articles for it to start really learning, it performs great with 1000 judgements and improves to around 8000 or so. There are techniques that could improve the “cold start” but they are not so interesting to me personally because I am swimming in judgements (although I could certainly subset the ones I have to do lots of experiments.)
There’s also a possible “pro” product which would be more configurable and aimed at say, professional patent searchers, recruiters, job seekers, people doing literature reviews and meta analysis and such.
There’s a definite tension between adding the features I want and doing the work that (I imagine) would make it useful for others, over time I am working on the architecture so I could put the features that I don’t want to share with others (like a mildly dangerous “engagement” model that I think many people would find annoying, social media share buttons for obscure platforms, etc.) into plugins and also get the data structures right so it can be forked w/o breaking the database.
I am using Superfeedr toingest RSS/Atom feeds into YOShInOn, my smart RSS reader, which costs 10¢ per feed per month, I am ingesting upwards of 2000 articles a day with a $8/month bill. I'd like to subscribe to 2000 small blogs that publish an article every week or month but that's not affordable even with bulk discounts and even if I stood up another ingestion system it would ridiculous to poll all those blogs 100 times for every article I get, and if more people started doing what I do, the load on those blogs would get unreasonable.
There used to be "Planet" software that would aggregate RSS feeds but the site for the software is dead although there are a few surviving planets like
https://planet.postgresql.org/
I wish we had more of them.