Case study · Built by Inseed
The LYVE App
The social commerce platform for Caribbean carnival culture: events, content, costume and ticket marketplace, mas band community, all in one app.
Problem
Caribbean carnival culture is large, passionate, and globally distributed, but the lived experience of it lives in five different places: Instagram for moments, WhatsApp groups for chatter, ticket sites for purchase, Facebook events for discovery, and word of mouth for the rest. There was no single place to capture the season, find what was happening, follow a mas band across countries, or buy and resell costumes and tickets.
Abraham wanted to fix that with one app on both iOS and Android, with real commerce attached. He needed it ready for App Store launch in months, not a year, and shipped to real users before the next carnival season.
Approach
We took the engagement end to end: Android and iOS apps, a public landing site, and an admin panel for Abraham and his team to manage carnivals, mas bands, listings, and users. The team peaked at 4 to 5 across mobile, backend, design, and QA. From day one we treated the App Store launch as a hard milestone and the first 1,000 users as the second one; everything was scoped against those two dates.
The build was a single codebase using React Native with Expo, paired with a NestJS backend and Supabase for Postgres, auth, and storage. WebSocket via the NestJS gateway powered chat. Payments split between Stripe for the marketplace (costumes, tickets, reseller flows, order history) and RevenueCat for in-app purchases. We invested early in Detox plus Cypress end-to-end testing because marketplace flows have to be reliable on launch day; chasing payment bugs in production is not a place we wanted to be.
We also stayed past the App Store launch. The brief explicitly included "support through the first 1,000 users", which meant performance tuning, App Store iteration cycles, and helping Abraham think through early marketing. That last part is outside what most agencies sign up for. We did it because the goal was a usable product on launch day plus 90, not a deliverable.
Stack
One decision worth telling
We started with NestJS exposing REST controllers. Two weeks in, the data fan-out was already a problem. The same costume listing was rendered in the seller dashboard, the buyer detail page, the cart, the order history view, and the carnival event page, and each surface needed a slightly different slice of the same record. With REST we were either over-fetching (sending the whole record everywhere and bloating the wire) or building parallel endpoints for each view (multiplying the API surface and the test matrix).
Moving to GraphQL resolvers, inside the same NestJS codebase, solved both. One schema, each surface declares exactly what it needs, no parallel endpoints. The migration cost about a week of backend work. Every subsequent feature landed faster, and the test matrix shrank instead of growing.
Outcome
“I honestly don't know where the LYVE App would be without Shuhel and the whole Inseed team. As soon as we started working together, it's been like magic, helping me bring what I've been dreaming about to life. And now the LYVE App has thousands of users.”
Video
Abraham, Founder, The LYVE App.
Selected screens
What we would do differently
Most agencies skip this section. We include it because honest hindsight is the kind of credibility you cannot buy.
We could have shaved 4 to 6 weeks off the first build by cutting features from the v1 scope and getting an MVP into the stores faster. The pre-launch period was thorough, but the customer feedback that started flowing in the first week after launch was more useful than any internal review.
The initial discovery feed surfaced five different content types (events, posts, mas bands, costumes, listings) in a single feed. Users got confused. We simplified post-launch, but a simpler v1 would have been a faster path to the same end.
Related work
Built by Inseed
Vibe search shipped in 3 weeks. AI searches up ~200%.
Read the case studyFounder-portfolio engagement
Mobile + light ML. Case study in progress.
Coming soonTell us what you are building. We will tell you what 4 weeks gets you, what 12 weeks gets you, and what we would do first.
Book your audit