Merewa AI is a cutting-edge, TikTok-inspired voice social platform that leverages Generative AI to create an immersive, personalized audio experience. Users can engage with a vertical, snap-scrolling feed of voice posts, assisted by AI personas that help generate scripts and respond to content in real-time.
- 📱 Immersive Feed: Smooth, vertical snap-scrolling experience optimized for voice content.
- 🤖 AI-Assisted Creation: Integrated LLM support (Groq/Ollama) for script generation and persona-based content creation.
- 🔍 RAG-Powered Search: Advanced vector search using Weaviate Cloud, enabling semantic discovery of posts and users.
- 🎙️ High-Quality Audio: Native voice recording and playback with visualized waveforms.
- 🔐 Secure Authentication: Multi-provider login (Google, GitHub, Email) powered by Better Auth and Supabase.
- ⚡ High Performance: Real-time caching with Upstash Redis and asynchronous processing with FastAPI.
- Framework: FastAPI (Python 3.11)
- Database: PostgreSQL (Supabase)
- Vector DB: Weaviate Cloud (RAG)
- Caching: Upstash Redis
- LLM: Groq (Llama 3.1) & Ollama
- Deployment: Render (Docker-based Web Service)
- Framework: Next.js 14 (App Router)
- Auth: Better Auth
- Styling: Vanilla CSS & Tailwind (Custom UI components)
- Deployment: Vercel
- Docker & Docker Compose
- Node.js 18+
- Python 3.11+
-
Clone the repository:
git clone https://github.com/mukesudo/Merewa-AI.git cd Merewa-AI -
Setup Backend:
cd backend cp .env.example .env # Update your .env with your Supabase/Groq keys docker-compose up --build
-
Setup Frontend:
cd ../frontend npm install npm run dev
Merewa AI follows a modern monorepo architecture:
/backend: Python FastAPI service handling the core logic, AI integrations, and database management./frontend: Next.js application providing a responsive, high-performance UI.
This project is licensed under the MIT License - see the LICENSE file for details.
Mukesudo - GitHub
Built with passion for the future of voice social media.