An offline-first language-learning platform that prepares learners for real conversations. The first course teaches Korean to English- and Vietnamese-speaking learners (A0 → practical A1, plus selected A2 travel/social scenarios).
After completing a level, the learner should be able to perform the real-world scenarios associated with that level.
- Scenario-based lessons (order food, ask directions, check into a hotel) instead of isolated drills
- Production-heavy exercises: learners speak and write from lesson one
- Works fully offline after content is downloaded; progress syncs safely later
- Natural speech: politeness levels, slang, texting conventions, and cultural context
- Personal error memory: the app remembers your recurring mistakes and reviews them
- Free core curriculum; AI is a supporting feature with deterministic fallbacks, never a dependency
vernora/
├── apps/ # User-facing applications
│ ├── mobile/ # React Native (Expo) learner app [planned]
│ └── admin/ # Next.js content administration [planned]
├── services/ # Backend services
│ └── api/ # Spring Boot modular monolith [planned]
├── packages/ # Shared code (API contract, content schema, design tokens)
├── infrastructure/ # Docker Compose, deployment scripts
└── docs/
├── architecture/ # System diagrams and design documents
└── decisions/ # Architecture Decision Records (ADRs)
The full product specification lives in vernora-description.md.
| Area | Technology |
|---|---|
| Mobile | TypeScript, React Native (Expo), SQLite, TanStack Query |
| Backend | Java 21, Spring Boot, PostgreSQL, Flyway |
| Admin | Next.js, TypeScript, Tailwind CSS |
| Platform | Supabase (Postgres/Auth/Storage), Docker, GitHub Actions |
| Speech (later) | C++20, whisper.cpp |
Pre-development. Currently building the product foundation (Phase 1 of the roadmap).