Samvaada AI is a professional-grade, government-standard helpline intelligence platform. It provides real-time multilingual voice transcription, automated intent analysis (NLU), and supervisor analytics for large-scale administrative operations.
- Real-time Transcription: Powered by local Whisper v3 models.
- Indic NLU: Intent and entity extraction for Indian languages using IndicBERT.
- Supervisor Dashboard: Live district-level sentiment heatmaps and operation KPIs.
- Secure Operator Portal: High-authority "CivicVoice" UI with 2FA simulation.
- 100% Local: Optimized for offline/air-gapped government environments.
/frontend: React + Vite + Tailwind v4 (UI Workstation)/backend: FastAPI + Python (Intelligence Engine)/ml-models: Local storage for Whisper and IndicBERT weights./scripts: Automation for model acquisition and system health.
- Python 3.10+
- Node.js 18+
- Virtual Environment (venv)
cd backend
# Create and activate venv
python -m venv venv
.\venv\Scripts\activate
# Install dependencies
pip install -r requirements.txt
# Download ML Models (First time only)
python ..\scripts\download_models.py
# Run server
uvicorn app.main:app --reload --port 8000cd frontend
npm install
npm run devOpen http://localhost:5176 to access the portal.
The project is configured for Vercel via the root vercel.json.
Important Note: Because this platform uses large-scale ML models (Whisper), the Backend Intelligence Engine requires a server with a GPU or significant RAM. For cloud hosting, it is recommended to host the backend on a GPU-enabled VPS or use a dedicated ML host (like Modal or Replicate) and point the frontend to that endpoint.
- Username:
admin - Password:
admin
© 2024 Samvaada Civic Intelligence | Built for Bharat