“Because every child deserves to be heard.”
🗣️ AI-powered speech companion for children with communication challenges.
Stem.me transforms speech therapy into an engaging AI experience for children aged 3–13 with speech impediments.
Using real-time voice recognition, AI-driven conversation, and scene-based storytelling, the platform encourages language growth in a fun, supportive way.
| 🌟 Feature | Description |
|---|---|
| 🗣️ Live Speech Recognition | Uses Web Speech API to understand children’s voice in real time. |
| 🤖 Conversational Therapy | Powered by Gemini 2.5 with adaptive correction for fragmented speech. |
| 🎨 Story Image Generation | Generates contextual scenery using Gemini Imagen 3 Fast for each conversation. |
| 🔊 Natural Voice Output | Uses ElevenLabs TTS for warm, child-friendly voice feedback. |
| 🌳 Progress Tracker | Visual growth tree rewards consistent speech and interaction. |
| Layer | Technology |
|---|---|
| Frontend | React + Vite + Web Speech API |
| Backend | Node.js + Express |
| AI Models | Google Gemini 2.5 (Text + Image), ElevenLabs (TTS) |
| Database | SQLite (default) / MySQL (optional) |
| Deployment | Netlify (Frontend) + Render/Heroku (Backend) |
git clone https://github.com/<your-username>/DivHacks-ChatBot.git
cd DivHacks-ChatBot/lingua