We are a non-profit organization that helps newsrooms, researchers, and public interest organizations host and share primary source documents. On the backend, we currently have 7.8 million documents, 131 million pages of searchable text, and 600 thousand individual shared notes in our database on production.
Each document has a somewhat complex permission system, which allows a document be set to private (only the owner can see it), shared (other members of their organization can see it), or public (publicly viewable and searchable), as well as a secondary set of permissions that allow documents to be added to project collections that also can grant access.
Users can put in a search on a centralized search bar that will parse through all of our documents and return matches that the user has permission to (if a person is not logged in, they can search through all public documents and get relevant results).