GitSearch attempts to do two things:
So let's take these in order.
If you've ever searched on GitHub or GitLab, it can be pretty obvious that finding public repos is not the primary thing those companies would like you to do. They are geared more towards the people writing the code than for people browsing. And perhaps that's how it should be.
Even so, it can be difficult to find basic sort and search options, and the UIs are burdened with not knowing whether you're trying to search Projects, Issues, Milestones, etc.
GitSearch makes this underserved use case into its main reason for being. When you have a technology or phrase that you want to query in just code repos (not webpages) then come to GitSearch and find all the stuff on GitHub and GitLab that you know and love, but also some cool stuff you can't find elsewhere.
As developers, we've seen the problems with monopolies and centralized services. Google can shut down any of its services arbitrarily and destroy entire businesses built off of them. Companies like Facebook can acquire and shutdown some of our favorite tools like Parse.
While at the moment it seems unlikely that Microsoft would shutdown GitHub, their monopoly (wink) on hosting code repos erases a lot of choice from the marketplace. There are many fine self-hostable GitHub alternatives, but if you decide to go elsewhere, it basically means your code is undiscoverable -- at least without a huge amount of press and developer adoption.
GitSearch will search any repo site with an API and public repos. Right now there is a manual submission process, but automating this is on the roadmap.