I don’t need a service, but I can see why a few people might. (For perhaps irrational reasons, I like services a teeny bit better, except perhaps for the weaker security, but @canton7, your arguments for a per-user tray app are good.)
Say there is a family of 3 that uses 2 computers: A, and B. User A1 shares some folders, and user A2 shares some folders. B is a laptop that is only on sometimes. When B turns on his/her computer, either the A1 folders or A2 folders (or both, if the computer is sitting at the login screen!) may be out of date and unavailable. This could be compounded if there is another laptop, C, that updates files in A1 or A2: B and C have no reliable hub to act as an intermediary for when they can’t both be online at the same time.
Bottom line: the availability of files suffers in some use cases with no central hub(s). (This could become elevated from an inconvenience to a severe problem if people are doing collaborative work on the same files.)
I have an always on Linux server, (that never has any user log into it), so I don’t care, but not everyone may have this option. If there is a service, then it might also be nice if the SyncThing security model was strong enough to prevent A1 and A2 from doing Bad Things™, or from having to share the same username/password, although families share username/passwords for things like home wifi routers.
But I am guessing there might be more high priority things to do for SyncTrayzor than trying to jam service support into there. People can still use NSSM (although if I understand correctly, they’d now be missing out on the file watcher functionality).