Talking about this stuff is starting to get the creative juices flowing again. I have been kicking around the idea of building a new budget app similar to YNAB4/Financier, using some of the newer C# technologies that have come out over the last few years. I work with that stuff at my job every day, so I would be very comfortable building something from scratch.
I have gone back and forth on whether it would be a simple, standalone desktop app, or a full blown web app (desktop & mobile) with full syncing capability. If I implement syncing, do I build it so it's self hosted for myself, and available for gurus to self host themselves, or set up a subscription service so anyone can sign up and pay for the syncing capability like Financier? If I didn't have to worry about hosting and securing data for other people, that would certainly keep things simple on my end. But if I build something new and useful, making it available and fully functional for everyone instead of just technology enthusiasts also sounds really attractive to me.
CouchDB provides really nice out-of-the-box syncing capability, but it also has some drawbacks when pushing the limits as we have encountered. It also makes it a little difficult to implement some features like scheduled transactions. I actually sat down and sketched out a plan for building a syncing algorithm a couple months ago that would solve some of those issues, but never got around to writing any code to test it out. I'm probably going to pull that sketch out again to see if I can put it into practice.
Financier has worked great for the last few years and I hate to reinvent the wheel, but it might be fun to build something new that solves some of the underlying issues and makes it possible to implement some of the features that have been requested like scheduled transactions, file import, search, etc.