EngReading
v5.0 — Korean sentence in, English brain out. AI classifies your verb into 1 of 7 Sets and animates how a native speaker builds the sentence left-to-right.
Forward-Flow English Brain Rewiring Engine. Input any Korean sentence, AI classifies the core verb into 7 Sets and animates the native English thought process. Gamified with XP, streaks, radar mastery charts, and SM-2 spaced repetition review.
The Backstory
The original EngReading was a “100-sentence story writing” platform — write one English sentence per day, get AI corrections, build a story over 3 months. It worked, but the client wanted something more fundamental: rewire the Korean brain to think in English word order.
The insight came from linguistics: Korean speakers fail not because they lack vocabulary, but because they assemble sentences backwards (SOV vs SVO). If you could teach the brain to predict what comes next in English — the way native speakers do — fluency follows naturally.
I rebuilt the entire platform from scratch around this idea. The result is v5.0: a Forward-Flow engine where you input any Korean sentence, AI classifies the core verb into 1 of 7 Sets, then animates how a native English brain would construct the sentence left-to-right, chunk by chunk. Every verb in English falls into one of 7 patterns, and each pattern predicts what comes next differently.
The 7-Set Verb Classification System
The core innovation. Every English verb belongs to one of 7 Sets, each with a distinct prediction pattern:
| Set | Name | Examples | Brain Predicts |
|---|---|---|---|
| 1 | Be Verbs | am, is, are | Complement (adjective/noun/place) |
| 2 | Action Verbs | eat, run, study | Object (what?) |
| 3 | Perception/Cognition | see, hear, feel | that-clause or object |
| 4 | Ditransitive/Causative | give, send, tell | Indirect + direct object |
| 5 | Linking/Change | become, get, turn | Complement describing state |
| 6 | Existence/Movement | go, come, arrive | Location/direction (where?) |
| 7 | Special Patterns | It is…that, What…is | Pattern-specific continuation |
Key Features
- Forward Flow Engine — Input Korean, see the English brain’s left-to-right thought process animated step by step
- 7-Set verb classification — AI identifies the core verb’s Set and shows its prediction pattern
- 6-step analysis — Verb extraction, Set identification, brain flow animation, final sentence, atomic chunks, KR vs EN comparison
- Quiz system — Multiple-choice verb prediction quiz after each analysis, awards XP
- Gamification — XP/level system (1-50), 5 tiers (Bronze to Diamond), 15 achievements
- 7-Set mastery radar — Recharts radar chart tracking proficiency across all 7 Sets
- Streak system — Daily streak counter with freeze tokens and streak bonuses
- SM-2 spaced repetition — Review past analyses at optimal intervals for retention
- Daily missions — AI-generated Korean sentences targeting weak Sets
- Dark luxury design — Gold accent (#C9A84C), glass morphism, shooting star animations
Technical Architecture
Next.js 16 App Router with Turbopack. Supabase for auth (cookie-based SSR), PostgreSQL with RLS, and service-role admin client for XP/achievement mutations. Fireworks AI (deepseek-v3p1) generates structured 6-step analysis JSON per Korean input. Framer Motion orchestrates staggered step-by-step animations. Recharts renders the 7-axis radar mastery chart. SM-2 algorithm schedules review intervals across 9 stages. Three Vercel cron jobs handle daily missions, streak settlement, and review scheduling.
What I Added Beyond the Brief
- Complete platform rebuild — v1-v4 was story writing; v5 is brain rewiring from scratch
- 7-Set classification system — Original linguistic framework mapping all English verbs to 7 prediction patterns
- Radar mastery tracking — Per-user, per-Set proficiency visualization with milestone badges
- Achievement system — 15 achievements across analysis count, streak, level, and Set mastery
- XP economy — Calibrated XP rewards for analysis, quiz, review, streaks, missions, and achievements
- Shipped to custom domain — engreading.com, live in production with Supabase backend
Screenshots
