A cost-efficient, multilingual voice assistant for budget Bluetooth speakers
Making smart voice assistance accessible to every Indian household
Smart voice assistants like Amazon Alexa and Google Home are powerful — but they require expensive proprietary hardware. Millions of Indians already own budget Bluetooth speakers from brands like boAt, Zebronics, Portronics, and more. These devices have no "smart" capability despite being perfectly capable audio output devices.
Why buy a ₹5000 smart speaker when you already have a ₹800 one?
Vani AI aims to be a cost-efficient, locally-aware voice assistant that:
- Works with any Bluetooth speaker — no new hardware needed
- Supports all major Indian languages (Hindi, Bengali, Tamil, Telugu, Marathi, Kannada, Gujarati, and more)
- Runs on a low-cost device (smartphone or Raspberry Pi)
- Provides region-aware responses tailored for Indian users
- Operates with minimal internet dependency for core features
| Feature | Description |
|---|---|
| 🎙️ Wake Word Detection | Custom wake word (e.g. "Hey Vani") |
| 🌐 Multilingual NLP | Understands and responds in 10+ Indian languages |
| 🔊 BT Speaker Integration | Seamless pairing with budget Bluetooth devices |
| 📡 Offline Mode | Core commands work without internet |
| 🏠 Smart Home Hooks | Control lights, fans via basic IoT integration |
| 🎵 Media Control | Play music, control volume, skip tracks |
| 📰 Local News & Weather | Region-specific updates in your language |
User Voice Input
↓
Wake Word Detection
↓
Speech-to-Text (Multilingual)
↓
Intent Recognition / NLP Engine
↓
Response Generator (Localized)
↓
Text-to-Speech → Bluetooth Speaker Output
- Speech Recognition: Whisper (OpenAI) / Vosk (offline)
- NLP / Intent: Rasa / custom model
- Backend: Java (Spring Boot) / Python
- TTS: Google TTS / Coqui TTS (offline)
- Hardware Target: Android phone / Raspberry Pi Zero 2W
- BT Communication: Android Bluetooth API / BlueZ (Linux)
- Finalize tech stack
- Build basic wake word detection prototype
- Implement Hindi + Bengali STT pipeline
- Add remaining Indian language support
- Integrate with Bluetooth audio output
- Offline mode for core commands
- Public beta release
This project is in early planning. Ideas, suggestions, and contributions are very welcome!
- Fork the repo
- Create a feature branch (
git checkout -b feature/your-idea) - Commit your changes
- Open a Pull Request
Subhojit Samanta — @Paradoxxt
"Voice assistance shouldn't be a luxury — it should be for everyone."
⭐ Star this repo if you believe in accessible AI for India!