WordUp: Native Spaced Repetition Vocabulary Engine
A native Android learning application for long-term memory retention. WordUp combines the FSRS-v4 memory scheduler with 62,500+ curated lexical entries. It runs 100% offline with zero cloud API costs.
Memory Science Without Cloud Fees or Latency
Rote memorization causes students to lose over 70% of new vocabulary within two days. Most commercial flashcard applications rely on simple interval counters or require costly cloud language models. These cloud solutions add network delay, privacy risks, and monthly operating expenses.
WordUp solves this problem with local algorithm execution. The application calculates individual memory stability, difficulty, and retrievability on the local device. It packages an indexed database of 62,502 words with phonetic data, collocations, and example sentences directly into the local application package.
Engineered for Rigorous Exam Prep and Professional Fluency
We designed WordUp to solve core structural limitations in traditional language software:
| User Persona | Traditional Problem | WordUp Engineering Solution | Result |
|---|---|---|---|
| IELTS & TOEFL Students | Unstructured word lists present low-utility terms. | Lexical data is indexed by CEFR proficiency levels (B2, C1, C2) and Zipf word frequency scores. | Learners master high-frequency academic vocabulary first. |
| Active Recall Learners | Standard flashcards only test passive recognition. | 17 drill formats test pronunciation, context clues, sentence composition, and collocations. | Students retain both active production and passive comprehension. |
| Commuters & Travelers | Applications stop working on flights, subways, and remote routes. | Pre-compiled SQLite database and native Android speech engines operate without network connections. | Daily study streaks continue without interruptions. |
| Privacy-Conscious Users | Ad networks and telemetry trackers collect personal learning data. | All progress records and quiz histories remain in private local device storage. | Zero personal telemetry leaves the device. |
The FSRS-v4 Mathematical Model
WordUp implements the Free Spaced Repetition Scheduler (FSRS-v4). This algorithm models human memory retention through three core variables:
- Stability (S): The time in days for recall probability to decline from 100% to 90%.
- Difficulty (D): The inherent complexity of the word on a scale from 1.0 to 10.0.
- Retrievability (R): The probability that the user recalls the word at the current timestamp.
┌────────────────────────────────────────────────────────────────────────┐
│ WordUp FSRS-v4 Memory State Flow │
└────────────────────────────────────────────────────────────────────────┘
New Word Presentation (State: NEW)
│
▼
First Review Rating (Again / Hard / Good / Easy)
│
├─► Again ──► State: LEARNING (Interval: Due Immediately)
└─► Good ──► State: REVIEW (Interval: Stability Formula)
│
▼
Subsequent Scheduled Reviews
│
├─► Correct Recall ──► Increases Stability (S) ──► Longer Interval
├─► Memory Lapse (0) ──► Resets Stability (S) ──► State: RELEARNING
│
▼
Leech Detection Engine
│
└─► 4+ Memory Lapses ──► Flags Word as Leech for Targeted Intervention
│
▼
Interval >= 21 Days ──► State: MASTERED (Permanent Long-Term Storage)
17 Modular Quiz Formats
The application generates daily quiz sets from overdue SRS items and new curriculum entries. Each format implements a modular Kotlin interface:
| Drill Format | Cognitive Mechanism | Verification Method |
|---|---|---|
| Classic Reveal | Flashcard recall with confidence ratings (Blanked, Sort Of, Knew It). | Direct FSRS rating mapping. |
| Sentence Lab | Sentence composition with target vocabulary in active context. | On-device TF-cosine and Jaccard semantic scoring. |
| Speak the Word | Verbal pronunciation practice with phonetic guidance. | Native Android SpeechRecognizer API. |
| Context Detective | Deduction of redacted words from full sentence context. | Distractor pool selection. |
| Collocations Drill | Identification of natural word pairings and idioms. | Curated collocation database match. |
| Word Family | Morphological root and derivation exploration. | Part-of-speech transformation rules. |
| Anagram Unscramble | Orthographic letter sequence reconstruction. | Character array permutation match. |
| Memory Match | Rapid association between words and definitions. | Paired matrix validation. |
| Rapid Fire & Speed Round | High-speed retrieval under strict time limits (10s to 60s). | Countdown timer with streak multipliers. |
High-Performance Android Engineering
Blackspring built WordUp as a modern native Android application. It adheres to clean architecture principles:
| Layer | Technology | Architectural Function |
|---|---|---|
| User Interface | Jetpack Compose | Declarative UI with high-contrast typography, smooth transitions, and zero dropped frames. |
| Local Persistence | Room (SQLite) | Stores 62,502 words and user records. Uses compound indexes for sub-3ms lookups. |
| Asynchronous Flow | Kotlin Coroutines & Flow | Reactive streams supply daily decks without blocking the main UI thread. |
| Dependency Injection | Dagger Hilt | Decouples database modules, repositories, and domain schedulers. |
| Semantic NLP Engine | Kotlin Vector Math | Calculates lexical similarity and validates sentences on-device without cloud models. |
| Speech Processing | Android SpeechRecognizer | Captures and checks spoken audio on-device. Eliminates speech API subscription costs. |
| Background Scheduling | Android AlarmManager & BootReceiver | Delivers daily study reminders reliably across device reboots. |
Gamification for Daily Study Consistency
Lex Mascot State Engine
An interactive companion character responds dynamically to study outcomes. Lex expresses distinct emotional states: thinking, celebrating, waiting, empathetic, and streak glow.
Flow State & Streak Tracking
The gamification engine calculates consecutive correct answers. It applies XP multipliers during sustained focus and alerts the learner when streak milestones are reached.
Memory Vault & Leech Management
Words that reach long intervals move into the permanent Memory Vault. Problematic words with repeated lapses receive automatic leech tags for targeted review.
Audited Technical Metrics
| Performance Metric | Industry Benchmark | WordUp Result | Operational Benefit |
|---|---|---|---|
| Daily Queue Generation | < 50ms | 2.8ms | Instant deck availability on application startup. |
| Database Query Latency | < 15ms | 1.2ms | Seamless search across 62,502 dictionary records. |
| UI Frame Rate | 60 FPS (16.6ms) | 60 FPS (0 dropped frames) | Smooth card gestures and tile animations. |
| Installed Package Footprint | < 50 MB | 23.4 MB | Includes complete dataset without post-install downloads. |
| Recurring Cloud Costs | $0.02 / user / day | $0.00 | Zero recurring API bills for operators or users. |
Build Custom Educational Systems with Blackspring
WordUp demonstrates Blackspring's capability in educational software engineering: offline data persistence, customized algorithmic scheduling, on-device linguistic evaluation, and high-performance native mobile interfaces. We build custom training platforms, compliance learning engines, and offline mobile tools for enterprises and institutions.