AI Voice Assistant | LLaMA 3.3 | Canvas Animation | Speech Recognition

🔥 **NAIRA - Voice AI Assistant with Real-Time Animated Avatar**

In this video, I build a complete voice-controlled AI assistant from scratch using:
- **Web Speech API** (speech recognition + synthesis)
- **Groq Cloud** (LLaMA 3.3 70B - 300 tokens/sec)
- **Canvas API** (animated face with blinking, mouth sync, particles)
- **Node.js** (simple proxy server)

✨ **Features:**
- Tap mic → speak naturally → AI responds by voice
- Living avatar reacts to: Listening → Thinking → Speaking
- Full chat history with context memory
- Voice selection (choose any system voice)
- Cyberpunk/neon UI with waveform visualization

📁 **Code:** [GitHub link]

🔑 **Get your free Groq API key:** console.groq.com

🕐 **Timestamps:**
0:00 - Demo (what NAIRA does)
1:30 - Architecture overview
3:15 - Backend proxy server (server.js)
6:00 - Frontend setup + Web Speech API
10:30 - Canvas avatar animation explained
15:00 - Groq API integration
18:00 - State management (listening/thinking/speaking)
21:00 - Styling + responsive design
24:00 - Final test
26:00 - Improvements & deployment

💡 **Learn:**
- How to proxy API requests without Express
- Real-time canvas animation loops
- Voice UX best practices
- Prompt engineering for spoken responses

**Subscribe for more AI projects!**

Видео AI Voice Assistant | LLaMA 3.3 | Canvas Animation | Speech Recognition канала U's-Solutions