This post is directed to the engineers. I was thinking about the struggles you face with building a file directory tree on someone’s google drive without having any of our files hosted on your servers.
I see the complexity where you have to rebuild a database of our google drive before you can start the sync process.
I think I just thought of a work-around that could be used as an option for some of your users.
The new Google Drive File Stream software has interfaces with the host OS that allow the OS to search and build indexes. Google stores this directory tree on their server and their software responds with it when the OS requests it. (Which can be demonstrated with how fast the cmd command 'tree' responds when run against GDFS’s mounted drive)
Knowing that Google Drive FIle Stream app mounts its folder as an internal disk.
Could your Insync software check if the official Google Drive app is already running and then just query that mounted disk for its directory structure?
This way you do not need to reinvent the wheel, just work as an “extension” or “companion” app to Google’s official programs. You could still use your old method of building the file database as a fallback.