Inspiration

Every night in Turkey, millions of parents ask the same question: "Did you study today?" And the answer is almost always the same: "yeah."

More than 3 million students sit LGS (high school entrance) or YKS (university placement) every year, and that one exam decides their next school. Families who can afford it hire a private coach. Everyone else plans alone and hopes it works.

We built Study Journal (Çalışma Defteri) to close the gap between study plans and study proof. Students already have timers and note apps; what's missing is an honest daily record that the student, the parent and the teacher can all trust, plus affordable coaching on top of it. Our goal: a study coach for every family, not just the ones who can pay for one.

What it does

Study Journal is a family-first study coach for Turkish exam prep, live on iOS and Android.

  • Students pick their exam and get a weekly plan built around their exam date. Logging a session takes about ten seconds (subject, minutes, questions solved) and works offline. Every log grows a topic map, so weak subjects show up early.
  • Stuck on a question: snap a photo and get a step-by-step explanation.
  • AI coach (Pro): pick a real university or high-school target, answer a short habit questionnaire, and get a personalized weekly plan. Every week the coach reviews the real logs, mock exam results and distance to the goal, and suggests concrete changes that apply with one tap.
  • Parents see the same day's progress, plan adherence, mock exam trends and monthly reports, without having to ask.
  • Teachers and coaches follow a whole class from one screen, see who hasn't studied in days, and share weekly programs with a link.

Everything is built around accountability: plan adherence, mock exam nets calculated with the official LGS and YKS rules, and weekly trends the whole family can see.

Business model

The person who pays is not the person who logs, so we priced around that:

  • Students: free. Logging, plans and the topic map stay free.
  • Teachers and coaches: free. They bring their whole class, which makes them our cheapest acquisition channel.
  • Parents pay for Family Pro (family_pro): one subscription covers every child in the house, with a 14-day free trial. ₺199.90/month or ₺1,490/year (~₺124/month, 38% off).

RevenueCat keeps one entitlement in sync across the whole family on iOS and Android, so a purchase on the parent's iPhone unlocks Pro on the child's Android phone.

Traction since launch

First public release on Google Play on August 2, 2026 and on the App Store on August 12, 2026, inside the Shipaton window. As of October 2, 2026 (production data, accounts created after the public release):

  • 175 accounts in 141 families
  • 13 teachers and coaches, some already running whole classes in the app
  • Growth is accelerating: 39 new accounts in the last two days alone (Sep 30 to Oct 1)

How we built it

A single Flutter codebase with Riverpod, GoRouter and Freezed models, built offline-first:

Layer Stack
Client Flutter, Hive (local cache)
Backend Firebase Auth, Firestore, Storage, FCM
Server Cloud Functions (push, RevenueCat webhook, AI)
Subscriptions RevenueCat → App Store + Google Play
AI coach and photo help Gemini via Cloud Functions
Engagement OneSignal + FCM push

Logs are written locally first and synced when online, which is critical for students on unreliable mobile networks. Families are modeled with invite codes, and role-based routing sends students, parents and teachers to their own home screens.

Pro is gated by a single entitlement (family_pro). A RevenueCat webhook keeps the family's billing state in Firestore, so access stays consistent across every device and platform in the household.

Challenges we ran into

  1. Offline-first sync. Study logs must never disappear. We designed Hive and Firestore reconciliation carefully, including conflicts when a student logs offline and syncs later.
  2. A Turkey-specific exam domain. LGS and YKS aren't generic "subjects." We built exam profiles, official date catalogs, multi-section mock exam templates (TYT, AYT SAY/EA/SÖZ), different net rules per exam and approximate score previews, while keeping the daily log fast.
  3. Family billing. One purchase must unlock Pro for the whole family, not just the paying parent's device. Wiring entitlements, store products (family_pro_monthly, family_pro_yearly), webhooks and Firestore billing state across iOS and Android was a multi-step puzzle.
  4. AI that earns its keep. We held the coach back until there was enough log data to review. It needs real placement targets, habit context and weekly thresholds, not generic study tips.
  5. Store readiness. App Store Server Notifications v2, Play billing service accounts, trial configuration and platform-specific RevenueCat keys all had to line up before we could test purchases end to end.

Accomplishments that we're proud of

  • Logging a study session takes about ten seconds, fully offline.
  • Parents see same-day progress without chasing their child for updates.
  • The AI coach loop works end to end: target selection → habit questions → generated weekly plan → weekly review with one-tap apply.
  • Stuck on a question: photo to step-by-step explanation.
  • Family Pro: one plan, every child, three roles (student, parent, teacher), 14-day trial.
  • Teachers are already using it: coaches run whole classes, send programs by link and follow adherence from one screen.
  • Shipped on iOS and Android from one Flutter codebase, with analytics on onboarding, paywall and retention.

What we learned

  • Accountability beats features. Users don't need another timer; they need an honest record that both student and parent trust.
  • The payer isn't the user. Making students and teachers free and charging the family turned teachers into a growth channel instead of a cost.
  • Offline-first is non-negotiable in mobile-first markets; sync has to be designed on day one.
  • RevenueCat let us focus on the family model instead of reinventing receipt validation and cross-platform entitlements.
  • Domain depth matters. In exam prep, localized data (placement scores, exam dates, net rules) matters as much as UI polish.
  • Gate AI on data. The coach is most valuable when it reviews real logs and mock trends, not on day one with an empty timeline.

What's next for Study Journal

  • School partnerships: bring whole classes in through teachers, the channel that is already working.
  • Smarter plans: adjust weekly load automatically after two weeks of adherence data.
  • Home screen widgets for quick logging and streaks.
  • Paywall experiments with RevenueCat offerings: trial length, yearly vs monthly default, family-size messaging.
  • Reach: ASO and exam-season content to bring affordable coaching to more of Turkey's 3M+ yearly LGS and YKS candidates.

Demo video: https://youtu.be/g2gZV8PSfvE

Built With

Share this project:

Updates

Submission history