Satark-AI: Defending Truth in the Age of Generative AI 🛡️

🌐 Live Demo

👉 Click Here to Open App

Satark-AI is a cutting-edge deepfake detection and speaker verification platform. It leverages advanced audio analysis techniques (MFCC, Spectral Analysis) and Deep Learning models to identify synthetic media and verify speaker identities in real-time.

---

📸 Screenshots

Dashboard	Mobile View

Live Monitor	Speaker Identity

---

🌟 Key Features

🕵️‍♂️ Deepfake Detection

• Advanced Audio Forensics: Analyzes spectral patterns and compression artifacts invisible to the human ear.
• Multi-Format Support: Upload MP3, WAV, or extract audio from MP4 videos.
• Explainable AI (XAI): Provides detailed confidence scores and heatmaps.

🆔 Voice Biometrics (Speaker Identity)

• Enrollment System: Create secure voice prints for known speakers.
• Verification: Match audio against enrolled identities with high precision using ECAPA-TDNN embeddings.

🎙️ Live Monitor

• Real-Time Protection: Continuously monitors microphone input (e.g., during calls).
• Chunk Analysis: Processes audio in 5-second intervals for instant feedback.
• Auto-Persistence: Saves all threats to a secure history database.

📱 Experience & Accessibility

• Progressive Web App (PWA): Installable on Mobile and Desktop.
• History & Playback: Review past scans, listen to audio, and download PDF reports.
• Smart Deduplication: Prevents redundant processing using SHA-256 file hashing.

---

🏗️ Architecture

The project is built as a highly scalable Monorepo using Turborepo:

• Frontend (apps/web): Next-gen UI built with React, Vite, Tailwind CSS, and Framer Motion.
• API Gateway (apps/api): Node.js with Hono framework for high-performance routing and Drizzle ORM for database management.
• AI Engine (apps/engine): Python service running FastAPI, PyTorch, and SpeechBrain for heavy inference.

---

🚀 Getting Started

Prerequisites

• Node.js (v18+)
• Python (3.10+)
• PostgreSQL

Installation

1. Clone the Repository

   git clone https://github.com/theunstopabble/Satark-AI.git
   cd Satark-AI

2. Install Dependencies

   npm install

3. Setup AI Engine

   cd apps/engine
   pip install -r requirements.txt

4. Configure Environment

   • Create .env files in apps/web (VITE_CLERK_KEY), apps/api (DATABASE_URL), and apps/engine.

Running Locally

Open 3 terminals to start the microservices:

1. Frontend:

   cd apps/web && npm run dev

2. API:

   cd apps/api && npm run dev

3. AI Engine:

   cd apps/engine
   # Activate venv if used
   python -m uvicorn main:app --reload --port 5500

---

🤝 Contributors

• Gautam Kumar - Lead Developer - LinkedIn | GitHub

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Made with ❤️ in India 🇮🇳