स्मृति · Smriti

Remember more. Study less. An offline-first spaced-repetition study app built for India's 30 million competitive-exam aspirants.

Submission for RevenueCat Shipaton 2026 — Next Gen Award (student track) Also entered for: RevenueCat Peace Prize · RevenueCat Design Award · HAMM Award

---

The problem

Every year around 13 million students sit for JEE, NEET and UPSC in India. Most of them study by re-reading notes — the single least effective revision method that cognitive science has measured. The students who do know about spaced repetition hit three walls:

1. Anki is brutal. Desktop-first, hostile UI, and $25 on iOS.
2. Data costs money. A student in a Tier-3 town cannot rely on a cloud app mid-revision.
3. No feedback loop. Nothing tells you "you will forget 40% of Polity by exam day."

Smriti fixes all three: a beautiful, fully-offline SRS app with a live Exam Day Forecast.

The core idea: Exam Day Forecast

Standard SRS apps show you what's due today. Smriti projects your memory forward to your actual exam date using a forgetting-curve model over every card's stability:

retention(card, examDate) = exp( −daysUntilExam / (stability · 2.2 + 1) )
forecast = mean(retention) across all studied cards

That single number — "you'll walk in remembering 74%" — is the app's whole motivational engine. It moves when you review, and students can watch it climb.

Features

🧠	SM-2 derived SRS engine tuned for exam cramming windows (shortened early intervals, 365-day cap)
📈	Exam Day Forecast — forgetting-curve projection to your real exam date
📴	100% offline — every card, every review, zero network calls. Built for patchy Indian data
🎴	3D flip review with 3-point rating (Forgot / Struggled / Mastered)
✍️	Notes → cards — paste notes, get a deck. Runs fully on-device, no API key, no internet
✨	Optional AI mode — bring your own free Gemini key for sharper cards from messy prose
📚	Pre-loaded decks — Modern Physics, Chemical Bonding, Indian Polity, Human Physiology, Calculus
🔥	Streaks & XP with per-card retention bars
⚡	Smriti Pro via RevenueCat — unlimited decks, AI cards, analytics, sync

Monetization (RevenueCat)

Freemium with limits that bind exactly when the app has already proven value:

• Free: 5 starter decks, 3 custom decks, 30 reviews/day, 3 card generations/day
• Pro: ₹99/mo · ₹699/yr (41% off) · ₹1,499 lifetime

Three packages ship in one offering so pricing can be A/B tested from the RevenueCat dashboard without a release. The paywall triggers contextually — at the daily review cap mid-session, when a student is motivated, not at cold launch.

src/lib/purchases.ts is the single integration boundary: configure, getOfferings, purchasePackage, restorePurchases, onEntitlementChange. On a native dev-client build these delegate to react-native-purchases against the pro entitlement; on web/preview they resolve locally so the complete paywall → purchase → entitlement-unlock loop stays demonstrable to judges.

Tech

• Expo SDK 54 + expo-router (file-based routing)
• React Native — one codebase, iOS + Android + Web
• TypeScript, strict, zero tsc errors
• react-native-svg for the animated forecast ring
• AsyncStorage — local-first persistence, no backend, no server bill
• RevenueCat for subscriptions

No backend by design: it keeps the app free to run and genuinely offline.

Run it

npm install
npx expo start        # scan the QR with Expo Go
npx expo start --web  # or run in a browser

Structure

app/
  _layout.tsx      store provider + dark stack navigator
  index.tsx        home — forecast ring, due count, deck list
  review.tsx       flip-card review session + rating
  generate.tsx     notes → cards, with a review-and-select step
  paywall.tsx      RevenueCat paywall (3 packages)
  new-deck.tsx     deck creator
  settings.tsx     progress, exam date, subscription
  deck/[id].tsx    deck detail + card editor
src/lib/
  ai.ts            offline note→card parser + optional Gemini engine
  srs.ts           SM-2 engine, retention & forecast maths
  store.tsx        app state, persistence, streaks, XP
  purchases.ts     RevenueCat boundary
  seed.ts          starter decks (JEE / NEET / UPSC)
  theme.ts         design tokens
src/components/UI.tsx

Why this deserves the Peace Prize

Access to good revision tooling is a quiet form of inequality. A student in Kota with a coaching subscription gets structured revision; a student in a village with a ₹5,000 phone and 1GB/day gets a PDF. Smriti works fully offline, ships with real syllabus content, and gives away the entire core loop for free. The paid tier subsidises the free one.

Licence

MIT — see LICENSE. Open source, as required by the Next Gen Award.