Raptures and RapKids, two comics apps from the first commit

Context

Raptures is a comics and animation platform for stories of Black American, African and Arab origin. I joined in October 2021 as the Flutter engineer and wrote the mobile app from its first commit. RapKids came later, a second app for children aged 4 to 12, and I opened that repo myself in October 2023.

This is private-company work, so there is no code to link. The two apps are in both stores, and that is the evidence.

Problem

A comics app looks simple from the outside: a reader and a list. Underneath it is a content business. Episodes have to load fast on cheap phones, purchases have to settle correctly on three different payment rails, and the team has to be able to ship a change without waiting a week for store review.

Then the company wanted a second product for a younger audience, on a different timeline, without slowing the first one down.

Constraints

Nigerian and diaspora readers on a wide range of Android phones, many on slow connections. Image weight and offline reading mattered more than animation.

Money on three rails: card payments through two processors and in-app purchases through the stores. Each one reports differently.

A small team. For most of the period the mobile group was under ten people, and I was the one who had to be able to explain any part of either app.

Decisions and trade-offs

Money settles from the provider, not from the client (throughout, with the last fix in August 2025). A purchase is not a purchase because the app says so. Subscriptions and coin purchases are confirmed from the payment providers’ webhooks, and the client only reflects what the backend has confirmed. That meant more backend work and a slower “success” screen. It also meant a failed webhook was a bug I could find and fix, rather than a customer who paid and got nothing.

Releases have to be reversible without a store submission (2024). I put remote-config feature flags and a force-update gate into both apps. A bad feature could be turned off in minutes. A build with a serious bug could be blocked from running. The cost is one more thing to check before every release. The benefit is that a store review queue stopped being an incident timeline.

RapKids got a different architecture on purpose (October 2023). Raptures used the MVVM and service-locator layout that made sense in 2021. For RapKids I chose a feature-first structure with BLoC, dependency injection and a declarative router. Two apps with two shapes is a real cost for a small team. I took it because the newer shape was easier to test, easier to hand to another engineer, and it was the pattern I wanted the team to grow into. I restructured the app flow once, in February 2024, and did the iOS submission preparation myself in July 2024.

Alert noise is a decision, not weather (August 2025). By 2025 the error tracker on Raptures had become a channel people ignored. I cleaned it up rather than letting it stay loud. A quiet channel that fires is worth more than a busy one nobody reads.

Give the content team their own tools (July 2024). Comic and episode image upload, delete and pagination in the back office, so editors could manage episodes without an engineer in the loop.

What was hard

Moving from writing the code to leading the people who write it. From April 2023 I was Engineering Manager across mobile, backend, web, DevOps and QA, a team of eight, while still being the person who knew the Raptures codebase best. The honest version is that I kept committing to both apps through the whole period. What changed was what I committed: releases, flags, payment paths and tooling, and less of the feature work.

Predictable releases. When I started, releases happened when things felt ready. Sprint rituals, code review, release management and an incident routine moved the team to a cadence people could plan around. None of that is clever. All of it had to be kept up every week.

Result

Raptures has been in both stores since February 2022 and is still shipping. On Google Play it shows 50K+ downloads and a 3.5-star rating as of September 2026.

RapKids went from an empty repo in October 2023 to both stores by July 2024.

I left the Engineering Manager role in September 2025 with both apps live, releases reversible, money reconciled from the providers, and a team that shipped on a schedule.