vigneshprasad/vismaya-kalike-voice

# Vika Voice Agent An outbound voice agent system for daily reflections with Vismaya Kalike facilitators. The system uses AI-powered voice calls to capture authentic stories and insights from learning center facilitators, prioritizing natural storytelling over structured reporting. ## Features - 🗣️ **Multi-language Support**: Kannada, Hindi, Tamil, and English - 🤖 **Multiple LLM Options**: GPT-4 and Gemini for conversational AI - 📞 **Automated Daily Calls**: Scheduled calls at facilitator's preferred times - 🎙️ **Audio Recording**: Full conversation recordings stored in AWS S3 - 📝 **AI Transcription**: Automatic transcription with Deepgram - 🧠 **Insight Extraction**: AI-powered analysis of themes, emotions, and stories - 🔒 **Privacy-First**: Automatic anonymization of children's names - 📊 **Data Export**: Easy export of reflections for research analysis ## Architecture ``` Voice Infrastructure: Vapi + Twilio (phone calls) Deepgram (speech-to-text) Azure/Google TTS (text-to-speech) Backend: Node.js + TypeScript + Express PostgreSQL (metadata & summaries) AWS S3 (audio recordings) AI Services: OpenAI GPT-4 / Google Gemini Native multilingual support ``` ## Prerequisites - Node.js 18+ with pnpm - PostgreSQL 15+ - Vapi account with API key - AWS account with S3 bucket - OpenAI API key (for GPT-4) - Google AI API key (for Gemini, optional) ## Quick Start ### 1. Clone and Install ```bash git clone cd vika_voice_agent pnpm install ``` ### 2. Environment Setup Create a `.env` file based on `.env.example`: ```bash cp .env.example .env ``` Fill in the required environment variables: ```env # Database DATABASE_URL=postgresql://user:password@localhost:5432/vika_voice_agent # Vapi (get from dashboard.vapi.ai) VAPI_API_KEY=your_vapi_api_key VAPI_PHONE_NUMBER_ID=your_vapi_phone_number_id # OpenAI (for GPT-4) OPENAI_API_KEY=your_openai_api_key # AWS S3 AWS_REGION=us-east-1 AWS_ACCESS_KEY_ID=your_aws_access_key AWS_SECRET_AC …