Why I stopped building separate backends.
Maintaining a separate backend made sense in 2018. In 2026 it's a tax most solo builders can't afford. Here's the moment the math flipped — and the stack that replaced it.
April 12, 2026
For the first few years of shipping software I did what every blog post told me to do. Frontend in one repo, backend in another. Database somewhere else. Auth bolted on as a third service. Deploy pipelines for each piece.
It worked. It also slowly killed every solo project I started.
The math is brutal once you write it down. Two repos means two deploy paths, two sets of secrets, two type systems that need to agree, two failure modes for every feature. As a solo builder, you don't have a backend team and a frontend team. You have *you* — context-switching between them with every change.
I stopped pretending that was sustainable. Now I use a single backend that lives next to the frontend. One type system. One deploy. One mental model.
The output: I ship apps in weekends instead of quarters. The features I actually wanted to build — real-time state, scheduled jobs, file uploads — show up in hours, not sprints.
That's the trade. Less choice, more shipping. If you're solo, I think it's the right one.