PractiPrep Code Sample

A sanitized TypeScript code sample inspired by the architecture behind PractiPrep, a mobile learning app for IHK Fachwirt exam preparation.

This repository demonstrates how premium entitlement state and AI usage limits can be resolved in a mobile app without exposing production source code, API keys, database schemas, AI prompts, payment product IDs, or proprietary business logic.

Why This Sample Exists

PractiPrep combines mobile learning flows, subscription access, AI-assisted explanations, account state, and backend synchronization. One of the most important product problems is deciding whether a user should receive access to premium AI features at a specific moment.

This sample focuses on that decision boundary:

1. Normalize entitlement signals from multiple sources.
2. Resolve one trusted premium access state.
3. Decide whether the AI assistant can be used right now.
4. Return a clear decision that the UI can act on.

What It Shows

• Type-safe domain modeling in TypeScript
• Subscription and offer-code edge cases
• Defensive handling of stale or unreachable backend state
• Daily AI usage limits for premium users
• UI-ready access decisions with reason codes
• Focused unit tests for product-critical logic

What Is Not Included

This is not the PractiPrep production app. It intentionally excludes:

• Real React Native screens or app navigation
• Real RevenueCat, StoreKit, or App Store product identifiers
• Supabase URLs, schemas, policies, functions, or service keys
• AI prompts, moderation rules, or provider credentials
• Question databases, answers, explanations, or learning content
• Private business logic and production implementation details

Core Flow

const entitlement = resolveEntitlement(sourceState, now);

const decision = canUseAiAssistant({
  entitlement,
  dailyUsage,
  now,
  policy,
});

The UI can then use the decision to either open the AI assistant, show a premium prompt, show a restore message, or explain that a daily limit has been reached.

Reviewer Guide

The most relevant files to review are:

• src/entitlement/resolveEntitlement.ts — normalizes purchase, restore, backend, and offer-code signals into one app-facing premium state.
• src/ai-usage/canUseAiAssistant.ts — turns entitlement and daily usage state into a UI-ready AI access decision.
• src/entitlement/resolveEntitlement.test.ts and src/ai-usage/canUseAiAssistant.test.ts — cover the edge cases that matter most for product behavior.

Repository Structure

src/
├── entitlement/
│   ├── entitlement.types.ts
│   ├── resolveEntitlement.ts
│   └── resolveEntitlement.test.ts
├── ai-usage/
│   ├── aiUsage.types.ts
│   ├── canUseAiAssistant.ts
│   └── canUseAiAssistant.test.ts
├── shared/
│   └── time.ts
└── index.ts

Run Locally

npm install
npm run check

Design Notes

The sample separates signals from decisions. Raw purchase, restore, offer, and backend states are treated as signals. The app-facing code receives a normalized entitlement object with a stable reason code.

That keeps screens simpler: UI components do not need to know whether access came from an active subscription, a recent purchase confirmation, an offer code, or a trusted cached state.

Public Scope

This repository is safe to publish because it contains only generic, sanitized TypeScript. It is intended as a readable engineering sample for portfolio and review purposes.

Author

Created by Nico Ohm / NOQCode.