Skip to content

SUBMISSION_DETAILS.md

Latest commit

 

History

History
53 lines (44 loc) · 2.65 KB

SUBMISSION_DETAILS.md

File metadata and controls

53 lines (44 loc) · 2.65 KB
Raw

🚀 Project Submission: Samvaada AI | CivicVoice Intelligence

📌 Basic Information

  • Title: Samvaada AI | CivicVoice Intelligence Platform
  • Synopsis: A professional-grade, government-standard helpline intelligence workstation featuring real-time multilingual ASR, NLU intent mapping, and Llama-3 powered response generation.

📝 Project Description

Samvaada AI is a comprehensive AI-driven helpline platform designed for the Karnataka Government (1092 Helpline). It transforms voice calls into actionable intelligence through a high-authority "CivicVoice" design system.

Key Capabilities:

  • Real-time Multilingual Transcription: Powered by Whisper-Tiny (optimized for low-latency).
  • Auto-Language Detection (LID): Automatically identifies Kannada, Hindi, and English speakers using the MMS-LID model.
  • Intelligence Loop: Integrated Llama-3-8B-Instruct to generate professional, empathetic responses in the caller's native language.
  • Supervisor Dashboard: Real-time operational analytics, sentiment heatmaps, and agent performance tracking.
  • Cloud-Hybrid Architecture: Optimized for Vercel deployment (<250MB) while maintaining high-fidelity AI performance via Hugging Face Inference API.

🔗 Project Links

🛠️ Instructions to Run

Local Setup (For Development/Privacy)

  1. Clone the Repo: 
    git clone https://github.com/Aditya-Ranjan1234/Samvaada-AI.git
cd samvaada-ai
  2. Setup Backend:
    • Create a virtual environment: python -m venv venv
    • Activate it: .\venv\Scripts\activate (Windows)
    • Install dependencies: pip install -r backend/requirements.txt
    • Run download script: python scripts/download_models.py
    • Start FastAPI: uvicorn backend.app.main:app --reload
  3. Setup Frontend:
    • cd frontend
    • npm install
    • npm run dev

Production Deployment (Vercel)

  1. Push the code to a GitHub repository.
  2. Link the repository to Vercel.
  3. Add the following Environment Variables in Vercel:
    • POSTGRES_URL: Your Supabase/Postgres connection string.
    • HF_TOKEN: Your Hugging Face API token (for Cloud ASR/LLM).
  4. Vercel will auto-detect the vercel.json and deploy both the React frontend and FastAPI backend.

🎯 Themes & Impact

  • Theme: AI for Governance / Civic Intelligence.
  • Idea: Shortlisted prototype for "Voice-based Multilingual Helpline Intelligence for Rural Bharat."

Generated for AI for Bharat Challenge 2026