Desktop English conversation practice app — Listen, Speak, get AI feedback.
npm installCopy .env.example to .env and add your key:
GEMINI_API_KEY=your_key_here
Get a free key at: https://aistudio.google.com/app/apikey
npm run devThis starts both the Vite dev server and Electron simultaneously.
npm startOr double-click the run.bat file. This builds the frontend for production and launches the Electron application locally under production settings (NODE_ENV=production), loading files from dist/ directly.
npm run packageOr double-click the package-app.bat file.
Output → release/ folder.
| Step | Action |
|---|---|
| 1 | Select a Part / Unit from the left sidebar |
| 2 | Listen to the English expression (auto-plays) |
| 3 | Press 🎤 and repeat the expression |
| 4 | Get your score + AI feedback |
| 5 | Press Next or retry until score ≥ 70 |
- 🔊 Listen — Auto-plays TTS on every new item; press 🔊 to replay
- 🎤 Speak — Web Speech API captures your voice; Gemini scores it
- ⭐ Notebook — Bookmark any item; export to
.txt - 💡 Coach — Tap Coach for a full AI explanation of the expression
- 📊 Dashboard — Track mastery per unit and daily streak
ActiveEnglish/
├── electron/
│ ├── main.js ← Electron main process, loads .env, IPC handlers
│ └── preload.js ← Context bridge (secure API key + store access)
├── src/
│ ├── context/
│ │ └── AppContext.jsx ← Global state (progress, bookmarks, navigation)
│ ├── components/
│ │ ├── SideNav.jsx ← Part/Unit navigation sidebar
│ │ ├── MicButton.jsx ← Record button with waveform animation
│ │ ├── FeedbackPanel.jsx ← Score display + AI correction
│ │ └── CoachPanel.jsx ← Slide-in AI explanation panel
│ ├── pages/
│ │ ├── StudyPage.jsx ← Main learning screen
│ │ ├── NotebookPage.jsx ← Bookmarked expressions
│ │ └── Dashboard.jsx ← Progress & stats
│ ├── hooks/
│ │ ├── useTTS.js ← TTS playback
│ │ ├── useSpeechRecognition.js ← Web Speech API mic
│ │ └── useAIFeedback.js ← Gemini API (feedback + coach)
│ ├── utils/
│ │ ├── scorer.js ← Fuzzy match scoring (0–100)
│ │ ├── storage.js ← electron-store IPC wrapper
│ │ └── audioPlayer.js ← TTS primary; unit MP3 playback
│ └── data/
│ ├── content.json ← 140 expressions (4 Parts, 12 Units)
│ └── mp3/ ← Per-item MP3s (optional, future use)
├── data/
│ └── 01.mp3 … 12.mp3 ← Unit-level native speaker audio
├── .env ← GEMINI_API_KEY (create from .env.example)
├── package.json
└── vite.config.js
| Score | Label | Status |
|---|---|---|
| 90–100 | Perfect! | Mastered ✅ |
| 70–89 | Great! | Mastered ✅ |
| 40–69 | Close! | In Progress 🟡 |
| 0–39 | Try again | Not mastered ❌ |
An item is marked Mastered when you score ≥ 70%.
| Layer | Technology |
|---|---|
| Desktop | Electron 28 (Windows) |
| Frontend | React 18 + Vite 5 |
| Styling | Tailwind CSS 3 |
| AI Feedback | Google Gemini 1.5 Flash |
| Speech Recognition | Web Speech API |
| TTS | Web Speech API (SpeechSynthesis) |
| Storage | electron-store (local) |
| Content | 140 items in content.json |