Shipping a real MVP in a week with Supabase and Vercel
Hangout.in, by UnboxingBLR, went from idea to live product in about a week. That is only possible because of what we deliberately did not build - and Supabase plus Vercel let us not build a lot.
Hangout.in, a product by UnboxingBLR, is one of our favourite examples of what our one-week MVP development practice can actually deliver. Not a prototype, not a clickable mockup — a live product, with real users, that you could put in front of the world. The reason that timeline is possible has very little to do with working fast and almost everything to do with what you decide not to build.
The one-week stack
For an MVP on this timeline, our default is Supabase and Vercel, and the work on hangout.in is a clean example of why.
Supabase gives you a real Postgres database, authentication, file storage and auto-generated APIs on day one. The day-one version of most products needs all four of those, and all four are things you would otherwise spend the better part of your week building, badly, instead of building the product. Supabase is not a toy version of these things — it is Postgres underneath, which matters, because it means the MVP is not built on something you have to escape from later.
Vercel removes deployment as a thing you think about. You push, it is live, previews and rollbacks are free. For a one-week build, the time you would spend on deployment infrastructure is time you simply do not have, and Vercel gives it back to you.
The point of the stack is not that it is trendy. The point is that it lets a small team spend the week on the actual product instead of on the scaffolding around it.
What we deliberately did not build
A one-week MVP is a scoping exercise wearing an engineering costume. The hard conversations are all about subtraction:
- No custom auth. Supabase auth, as it comes. Every hour spent on a bespoke auth system in week one is an hour stolen from the thing that makes the product worth signing up for.
- No premature abstraction. We did not build for the scale we hoped for. We built for the scale we had, which was zero, and made sure the foundation — real Postgres — would not have to be thrown away when scale arrived.
- No admin panel, no settings sprawl, no "while we're here". The MVP did the core thing. Everything else waited for evidence that anyone wanted it.
This is the part that is genuinely hard, and it is hard for the client more than for us. Cutting scope feels like cutting ambition. It is the opposite — it is what gets the ambition in front of real people fast enough to learn something.
The pitfalls
The "we'll just add" creep. Every small addition in week one is cheap on its own and expensive together. We hold a hard line: if it is not core to the thing being validated, it is a week-two conversation. Most week-two conversations never need to happen, because the live product answered the question for free.
Treating the MVP stack as the forever stack — or as disposable. Both are mistakes. Supabase and Vercel are genuinely good and many products never need to leave them. But because Supabase is real Postgres, if a product does outgrow the managed setup, the data and the core do not need a rewrite — they need a migration, which is a much smaller, much more boring thing. We build the MVP knowing both outcomes are fine.
Skipping the unglamorous basics. Fast does not mean careless. Even in a week, the database schema gets thought about properly, because the schema is the part that is genuinely painful to change later. Row-level security is on. The boring foundations are exactly the things you do not get to skip, because they are the things you cannot cheaply fix afterwards.
What we would tell you
A one-week MVP is real and it is worth doing, but understand what makes it possible: a stack that hands you the common infrastructure for free, and the discipline to cut everything that is not the core idea.
- Pick a stack that removes scaffolding. Supabase and Vercel exist so you can spend the week on the product.
- Scope by subtraction. The question is not "what could it do", it is "what is the smallest thing that proves the idea".
- Do not skip the schema and the security. Those are the expensive-to-fix parts. Everything else can wait.
- Build on foundations you would not be ashamed to grow. Real Postgres underneath means the MVP is a starting point, not a liability.
Hangout.in shipped in about a week because of a hundred small decisions to not build something. That is the actual skill. The code was the easy part.
Working on something similar?
We have probably hit the exact scaling or architectural bottlenecks you are navigating. Book a call and we'll walk through the technical trade-offs.