meligavincent/Mveng

🌍 African AI assistant powered by Axum + Whisper + Ollama. Digital griot sharing African stories, history & culture. Rooted in Cameroon's Yesum traditions. Named after scholar Engelbert Mveng. Your bridge to African wisdom. 🇨🇲✨ # Mveng 🌍 African AI assistant powered by Axum + Whisper + Ollama. Digital griot sharing African stories, history & culture. Rooted in Cameroon's Yesum traditions. Named after scholar Engelbert Mveng. Your bridge to African wisdom. 🇨🇲✨ # VoiceAI Assistant 🎙️🤖 A conversational AI agent that processes voice input, transcribes it using OpenAI Whisper, generates responses with Ollama (open-source LLM), and provides both text and optional voice responses. ## 🚀 Features - **Voice Input Processing**: Upload audio files for transcription - **Real-time Transcription**: Uses OpenAI Whisper for accurate speech-to-text - **Open Source LLM**: Powered by Ollama with Llama 2 or other open models - **RESTful API**: Built with Axum (Rust) for high performance - **Containerized**: Complete Docker Compose setup - **WebSocket Support**: Real-time conversation updates - **Multi-format Audio**: Supports WAV, MP3, M4A, and more - **Response Streaming**: Stream LLM responses in real-time ## 🏗️ Architecture ``` ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ Client App │───▶│ Axum Server │───▶│ Whisper Model │ │ (Web/Mobile) │ │ (Rust) │ │ (Transcription)│ └─────────────────┘ └─────────────────┘ └─────────────────┘ │ ▼ ┌─────────────────┐ │ Ollama Server │ │ (Open Source LLM)│ └─────────────────┘ ``` ## 📋 Prerequisites - Docker and Docker Compose - At least 8GB RAM (for LLM models) - NVIDIA GPU (optional, for faster inference) ## 🛠️ Installation 1. **Clone the repository** ```bash git clone github.com cd voiceai-assistant ``` 2. **Start the services** ```bash docker-compose up -d ``` 3. **Pull the LLM model** (first run only) ```bash docker-compose exec ollama ollama pull llama2 ``` 4. **Verify installation** ```bash curl localhost ``` ## 🐳 Docker Services ### Core Services - **axum-server**: Main API server (Port 3000) - **whisper-service**: Speech-to-text …