Back in the late 90s we had a publishing system called Manila. This was around the time Google was starting, and they weren't yet indexing blogs. We needed to have search working for our sites, so we wrote a search engine that ran inside Manila. It worked the way "real" search engines do. It crawled the pages and created an index. So it didn't have to do any actual searching when you queried it. 

Eventually we came to depend on Google, but now in 2016 Google is not very good at finding stuff in my archives. And it seems to be getting worse. Phrases I used many times over the years come up with zero hits. I know how to use Google, it became part of my publishing system. I still use it for lots of stuff, but it's getting ineffective at searching my blog.

So all this leads me to the question -- has anyone made a simple easy to set up search engine for Node? It should crawl sites over HTTP. It should have a ranking algorithm, so it returns more relevant hits first. Launch the app in Forever and use a browser-based interface to tell it what to crawl. Access the search engine through the web, of course.  

If it doesn't exist, we should create it. We need to have an existence independent of Google. Strategy taxes on our web, not good. 

Update: Simple-search.