In general, bare repository is the one in which there's no working directory. Instead, there's metadata (a specific database) included.
It's good to know that the main difference between bare and regular repository is that the regular one has both metadata and working directory included. In such repository it's possible to browse, change and commit files (you cannot to that in a bare repository).
You may ask: Hey, what's the goal of using bare repository in central repositories if it's basically the same and the only thing is that working directory is cut?
The thing is that bare repository is usually used in central repositories (you can either push to it or pull from it).
Important: In central repositories, however, you shouldn't use the repositories with a working copy since such approach may cause your data loss