Speak the Language of Your People. Empower your business with an AI voice assistant that understands local dialects and nuances.
An AI voice assistant built for small businesses in Africa, helping them respond to customer voice or text queries in multiple native languages (Pidgin, Yoruba, Swahili, and 10+ more).
We are proudly building VOXY as part of the African Tech Journal Hackathon.
Follow our journey as we ship, learn, and grow in public.
#BuildInPublicATJ
Small businesses in emerging markets often struggle to manage customer inquiries, orders, and support during peak hours. Furthermore, standard AI solutions struggle with local accents, dialects, and code-switching (e.g., blending English with Pidgin or local languages), leading to frustrating customer experiences.
Voxy is an intelligent voice and text assistant tailored specifically for African businesses. It deeply understands native accents—not just translating, but truly comprehending local context. It seamlessly handles inquiries, takes orders, and provides support, ensuring no customer is left waiting.
- Native Accents & Dialects: Deeply trained to understand and speak Pidgin, Yoruba, Swahili, and more.
- Lightning Fast: Optimized response times (< 1s) to keep customers engaged without awkward pauses.
- Privacy First: Local processing options available for sensitive business data.
- Omnichannel Support: Handles both voice notes and text messages.
- Business Dashboard: Track conversations, manage business settings, and view AI performance analytics.
Frontend:
- Next.js (React 19)
- Tailwind CSS v4
- Framer Motion (Implicit/Tailwind Animations)
Backend:
- Node.js & Express.js
- PostgreSQL (pg)
- RESTful API Architecture
- Customer Interaction: A customer sends a voice note or text in their local dialect.
- Processing: The audio/text is sent to our backend where the AI transcribes and understands the intent, accounting for local slang and context.
- Action: The AI queries the business's specific knowledge base or inventory.
- Response: A natural, culturally-aware response is generated and sent back via text or synthesized localized voice.
- Node.js (v18+)
- PostgreSQL Database
- NPM or Yarn
git clone https://github.com/your-username/voxy.git
cd voxycd server
npm installCreate a .env file in the server directory and add your environment variables:
PORT=5000
DATABASE_URL=your_postgres_connection_string
# Add other necessary API keys (OpenAI, ElevenLabs, etc.)Start the backend server:
npm run devOpen a new terminal and navigate to the root directory:
# from the root 'voxy' directory
npm installStart the Next.js development server:
npm run devVisit http://localhost:3000 to view the application.
- Integration with WhatsApp Business API to handle messages directly in-app.
- Expanding language support to 20+ African languages.
- Voice cloning options for business owners to use their own voice.
- Automated inventory management syncing based on customer orders.
Built with ❤️ for the African Tech Journal Hackathon.