AI · Education · Voice
Word Fun
Word Fun is an AI-powered phonogram-based reading app for children in Kindergarten to Grade 1 (ages 5–6).
It listens to a child pronounce a word, analyses their speech in real time using the OpenAI Whisper API,
and provides instant audio feedback — ideal for pre-readers who cannot yet read instructional text.
The app breaks words into phonograms (e.g. “school” → sch, ool) and rewards correct
pronunciation with stars, unlocking a bubble-pop game at milestone counts.
It was built and deployed independently by Nick Abbott.
Key Features
- Listens to a child pronounce a word and gives instant feedback using the OpenAI Whisper API.
- Breaks words into phonograms with audio help — e.g. “school” → sch, ool — spoken aloud via the browser's Web Speech API.
- Star-based reward system: 3 stars for unprompted success, fewer for attempts with help.
- At 10 stars, the app automatically unlocks the Puppy Fetch game with a celebratory announcement.
- Audio-first design — all feedback is spoken, making the app fully accessible to pre-readers.
Technology Stack
| Layer | Technology |
|---|---|
| Backend language | Java 25 (JVM) |
| Backend framework | Jakarta EE / JAX-RS on WildFly 39 |
| Database | H2 (embedded) |
| AI — speech recognition | OpenAI Whisper API |
| Text-to-speech | Web Speech API (browser built-in) |
| Frontend | React 19 + Vite 7 (JSX) |
| Build | Maven — WAR deployed to WildFly |
| Deployment | Docker container on AWS (ECS Fargate) |