I'm a solo founder from Denmark, and I'd describe myself as non-technical, my background is design, not engineering. A few weeks ago I shipped an iOS app called VibeFeed, and it's since crept into the top of the Danish App Store's Social Networking category. This is the honest writeup: why I built it, the stack that let one non engineer ship it, and what I got wrong.
The itch
Every feed I open has slowly turned into the same thing: ads, strangers, and an algorithm deciding what I see. The handful of people I actually care about get buried under all of it. I wanted the opposite, a quiet, chronological feed shared only with the small circle of people I'd actually invite to dinner. No ranking. No infinite scroll. No public like counts. No follower numbers.
That constraint, deliberately small, became the whole product.
The stack
I'm not going to pretend I hand rolled a distributed backend. The reason a non-technical founder can ship a real app in 2026 is that the tooling does the heavy lifting:
Next.js on Vercel for the web side and deploys. Push to GitHub, it's live. No servers to babysit.
Supabase as the backend — Postgres, auth, and row-level security. RLS turned out to be the single most important piece, because the entire premise of the app is "you only see posts from your own circle." That's not a UI rule, it's a data-access rule, and Supabase's policies let me enforce it at the database level instead of hoping my frontend never leaks.
Stripe for payments.
The lesson that surprised me: the hard part wasn't the code, it was modeling the data so that "a private circle of friends" was true at the row level and couldn't be worked around. I redid that schema more times than I'd like to admit.
What I got wrong
Onboarding. A close, friends app is worthless to the first person through the door — they open it and the feed is empty, because their friends aren't there yet. I underestimated how much the onboarding has to pull your circle in before the app makes any sense. That's the thing I'm rebuilding now.
Trusting my frontend too early. My first version enforced privacy in the app layer. Moving that down into RLS policies was the moment I actually trusted the thing with people's photos.
Chasing the wrong number. Being non-technical, I fixated on install counts and store ranking early. For a social app that lives or dies on real friend-graphs, the only number that matters is whether a circle stays active a week later. Everything else is vanity.
Where it is now
It's early and honest: a couple dozen users, some of them still my test accounts, iOS only. The "calm / anti-algorithm social app" graveyard is deep — Path is the obvious ghost — and I'm well aware most of them died the moment they chased scale and bolted on growth-hacky mechanics. My bet is that keeping the circle genuinely small, forever, is the thing they got wrong.
If you've built something in the anti-algorithm / digital-wellbeing space, I'd genuinely like to hear what worked and what didn't — especially around onboarding and getting a whole friend-group in at once rather than one lonely user at a time.
Top comments (0)