Viktor Sokoliuk product engineer · EU
← All cases

Case · own product · EdTech

Learn Hebrew: one developer, two platforms, a subscription business.

How I took a Hebrew alphabet app from an Android side project to 10K+ downloads, a native iOS version and growing subscription revenue — without an ad budget or a team.

Google Play ↗App Store ↗hebrewedu.com ↗
10K+
downloads on Google Play
4.2★
from 109 reviews
3,200
active Android users
4×
ARR growth in 12 months
7.41%
iOS day-7 download-to-paid
My role
Everything: product, UX, Android, iOS, backend, content pipeline, ASO, analytics.
Platforms
Android, iPhone (2026), web: hebrewedu.com in English and French.
Stack
Kotlin, Jetpack Compose, Swift, SwiftUI, Django, RevenueCat, Firebase, Astro.
Business model
Freemium with daily limits. Premium $4.99/month or $29.99/year.

The challenge

Hebrew is hard for beginners for one reason first: a new alphabet, written right to left, mostly without vowels. Most apps jump to vocabulary and lose people at the letters.

The constraint was mine too: one developer, no investor money and no ad budget. Every feature had to earn its place by retention or revenue.

What I built

Alphabet course22 letters, final forms and niqqud, with native audio and letter tracing.
Reading trainerBeginner stories with audio, toggleable translation and comprehension quizzes.
Verb engineHundreds of verbs with roots, binyanim and full conjugation tables, offline.
AI word explanationsTap any word for letters, vowels, pronunciation and meaning, served by my own API.
Spaced repetitionDictionary and flashcards that bring back the words you keep missing.
Habit loopDaily drills, streaks, achievements and a well-timed review prompt.

How the pieces fit

Two native apps and a website share one backend and one subscription system, so content and purchases stay in sync.

Search traffic
hebrewedu.com
Astro · lessons, games, EN/FR
Android
Kotlin + Compose
Google Play Billing
iOS
Swift + SwiftUI
StoreKit · iPhone
Shared
api.hebrewedu.com
Django · RevenueCat · Firebase

Five decisions that moved the numbers

  1. 01Subscriptions over ads Removed interstitial ads entirely. Fewer interruptions, better reviews, and premium became the obvious upgrade.
  2. 02Native iOS, not a rewrite in a framework Ported the app to SwiftUI with a 106-task plan in five phases, checking each step against the Kotlin source. iOS users convert at 7.41% by day 7.
  3. 03Own backend for AI Moved AI word explanations from Cloud Functions to my Django server: predictable cost, no cold starts, one API for both apps.
  4. 04A content site as the funnel hebrewedu.com answers what beginners search for, with free browser games, then points them to the apps. Built from a keyword-clustered semantic core.
  5. 05Localize where demand is Store listings in Russian, French and German; the French site mirrors English page for page instead of a thin translation.
What didn't work (yet)

App Store search is harder than Google Play

On iOS a weaker competitor still outranks the app for "Learn Hebrew". Organic ASO alone is slow there; the next step is a measured Apple Search Ads test with a fixed budget.

What this means for your project

You get the playbook, not just the code

Paywall placement, onboarding, store listings, analytics events and a plan for the first 1,000 users — the parts agencies leave to you.

See MVP packages →

Building a learning or subscription app? Let's talk.

Book a 20-min call All cases →