DAPS — Digital Asset Protection System

A production-style full-stack platform for protecting digital assets using machine-learning fingerprinting, automated piracy detection, and AI-assisted enforcement workflows.

---

Overview

DAPS helps creators and rights holders register original media, generate unique fingerprints, monitor potential infringements, and respond with faster legal and operational workflows.

The platform combines:

• a React/Vite frontend for dashboards and analyst tooling
• a Node.js/Express API for authentication, orchestration, and persistence
• a FastAPI ML microservice for image/video fingerprinting and similarity checks
• a background scraper for identifying and reporting potential piracy events
• MongoDB for storing users, protected assets, and violations

---

Core Capabilities

• Responsive Cyberpunk UI with modern dark mode aesthetics and data-rich analyst toolsets
• Secure media registration with upload, storage, and ownership mapping
• Visual fingerprint generation using perceptual hashes and CNN embeddings
• Image/video comparison for duplicate and infringement detection
• Hybrid hit-and-run scraper (Cheerio + Puppeteer) for fast, memory-efficient background scanning and reporting
• Violation intelligence dashboard for reviewing detected breaches
• Robust AI Legal Assistant using a multi-model fallback system (OpenRouter) to generate legal strategies and copyright takedown requests
• ML sandbox for side-by-side diagnostic comparison during demos and testing

---

System Architecture

Layer	Stack	Responsibility
Frontend	React, Vite, Tailwind	Cyberpunk UI, auth views, ML sandbox, violation dashboard
Backend API	Node.js, Express, Mongoose	Auth, uploads, AI fallback routing, orchestration
ML Service	FastAPI, PyTorch, TorchVision, OpenCV	Fingerprinting, embedding generation, similarity scoring
Scraper Node	Node.js, Axios, Puppeteer, Cheerio	Hybrid hit-and-run asset scanning and reporting
Database	MongoDB Atlas / Docker	Users, media records, violation records

---

Repository Structure

GSS/
├── backend-node/      # Express API, auth, uploads, AI endpoints, MongoDB integration
├── backend-python/    # FastAPI ML service for hashing, embeddings, and comparisons
├── frontend/          # React frontend variant
├── frontend-react/    # Primary Vite/React UI used for dashboard and demo flows
├── node-scrapper/     # Background scraper and reporting pipeline
├── combined_architecture_flow.md
├── backend_architecture_flow.md
├── frontend_architecture_flow.md
├── python_architecture_flow.md
└── scraper_architecture_flow.md

---

End-to-End Workflow

1. User uploads an original asset through the frontend.
2. The Node backend accepts the file and sends it to the Python ML service.
3. The ML service generates a fingerprint profile using hashes and embeddings.
4. The backend stores the media and metadata in MongoDB.
5. The scraper service scans external sources for likely matches.
6. Suspect files are verified by the ML service and reported back as violations.
7. The frontend displays violations and exposes AI-assisted response tools.

---

Local Development Setup

Run with Docker Compose (Recommended)

To run the entire stack seamlessly and avoid local environment configuration issues, use the provided Docker Compose setup:

1. Configure Environment Variables: Ensure you have created the necessary .env files locally before running Docker.

   • backend-node/.env (See the Node backend .env example below)
   • node-scrapper/.env (Requires BACKEND_URL=http://backend-node:5000 and SCRAPER_API_KEY)

2. Start the Stack:

docker-compose up --build

This builds and starts the Frontend (accessible at http://localhost), Node API, Python ML Service, Scraper, and a local MongoDB container, automatically linking internal networks and environment variables.

Manual Setup (Without Docker)

If you prefer to run services individually:

Prerequisites

• Node.js 18+
• Python 3.10+
• MongoDB Atlas or local MongoDB
• npm / pip

1) Start the Node backend

cd backend-node
npm install
npm start

Example .env for backend-node/.env:

PORT=5000
JWT_SECRET=your_jwt_secret
MONGODB_URI=your_mongodb_connection_string
OPENROUTER_API_KEY=your_openrouter_api_key
SCRAPER_API_KEY=daps-internal-key-2026
PYTHON_SERVICE_URL=http://localhost:8000
BACKEND_URL=http://localhost:5000

2) Start the Python ML service

cd backend-python
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
python run.py

3) Start the frontend

cd frontend-react
npm install
npm run dev

4) Optional: run the scraper

cd node-scrapper
npm install
node index.js --run-now

Local service URLs

• Frontend: http://localhost:5173
• Backend API: http://localhost:5000
• Python ML Service: http://localhost:8000

---

Major API Groups

Authentication

• POST /api/auth/login
• POST /api/auth/register
• POST /api/auth/sendotp
• POST /api/auth/forgot-password
• POST /api/auth/reset-password

Media

• POST /api/media/upload
• GET /api/media/me
• GET /api/media/violations/me
• GET /api/media/all
• POST /api/media/report

AI Assistance

• POST /api/ai/analyze
• POST /api/ai/summarize
• POST /api/ai/draft-takedown

ML Service

• POST /compare/image
• POST /compare/video
• POST /compare/two-images

---

Deployment Note

The project can be deployed to services such as Render, but the Python ML service may be noticeably slower on free-tier hosting because:

• the service can sleep and cold-start after inactivity
• torch / torchvision model loading is heavier in cloud CPU environments
• fingerprint generation happens during real-time upload and comparison flows

For smooth live demonstrations, the local environment provides the best real-time experience.

---

Project Highlights

• Clean separation of concerns via microservice-style architecture
• Real-world combination of security, computer vision, and AI-assisted legal tooling
• Designed for demo readiness, experimentation, and future productization
• Extensible structure for adding new detectors, models, or enforcement workflows

---

Status

This repository is an advanced prototype / demo system for digital asset protection and piracy monitoring. It is ideal for showcasing:

• full-stack system design
• ML service integration
• media fingerprinting workflows
• AI-enhanced operational tooling
• multi-service deployment patterns

---

License

This project is intended for educational, prototype, and demonstration use unless otherwise specified by the repository owner.