I’m a little late here but I can explain it.

What Sonarr, Radarr and the others do automatically is scan an RSS feed of new releases. Basically, they look through the things that were uploaded to each tracker, since the last scan, to see if anything should be downloaded. The issue with this is that this only finds the things that were uploaded since the media item (show or movie) was added, not things that were uploaded before then. What this software does (and Huntarr used to do) is run a full search on the indexers, so the *arr apps can find any file that has ever been uploaded to the trackers. It’s the same search you can trigger manually.

Technically, adding a media item, manually triggering a full search once and then relying on the automatic RSS scan should give you the same result (unless any trackers were down while you ran the manual search, which is one of the issues with this approach) but, for example, I use Seerr, so friends and family can add new stuff to my *arr apps but Seerr doesn’t trigger a full search when something gets added, so I would have to manually trigger the full search once every time someone adds something. The point of having something like Fetcharr is that the full search gets triggered automatically. In addition, it does so periodically instead of just once, so that fixes issues like trackers being down when the search was triggered or a download not being found in the RSS feed despite it existing (not sure if the second one ever happens but it technically could and this would solve that issue).

Edit: I also just saw that the GitHub page for this has an explanation as well:

I learned that at *arr stack uses RSS feeds to scan for and fetch updates, so if your indexer doesn’t support those feeds or the feeds or too old (or a myriad of other issues that can come from this kind of system) then you won’t get replacement content even if it exists. This is why these kinds of “hunting”, “fetching”, “upgrading”, etc systems work so well. They simply force the *arr apps to periodically update their content through their configured indexer, regardless of RSS feed availability.

Currently using zsh but I installed fish yesterday to try it out because I’m thinking of switching. All the zsh plugins I have are basically just replicating what fish has by default anyway and fish might do it better.