This all looks very interesting & exciting. But you all know I'm not a coder but I know enough to get me into trouble. But with my tin foil hat getting a whole lot more fancy and bigger lately I look at that fancy diagram and all I can think of is the presnetation of a multitude of attack vectors and points of failure in a very tightly dependent system.
You are going to need a much bigger developement team and a lot of ongoing maintainers.
And money.
I do hope you guys can pull it off and make it work. Thats one hell of an undertaking.
I wouldn't consider it all to be incredibly tied in a way that'll fail. The bulk of the work being done by this piece of software is separated out to numerous micro services. For example union indexer, encoder, ipfs-cluster, etc. Just the difference is we have a core backend that hooks all this stuff together compact and easy to run way.
On more sensitive topics like account creation/authentication, We (me) are looking at building a decentralized multisig to handle account creation and other ways to further decentralize/derisk our account creation proxy. Especially for proxy accounts where you can "select" a provider to be the account you ultimately post on chain... Of course, I am talking about further into a future. One step at a time right?
Baby steps are awesome as long as it isnt into a mud pit. LOL!