🛡️ ConsentIQ - Data Nutrition Label Intelligence

Transforming complex privacy legalese into readable, student-friendly data nutrition labels.

Built for the AMD Slingshot Hackathon, ConsentIQ is a privacy-first analysis tool that scans AI application policies, identifies high-risk data practices using local NLP, and generates a visual risk report—completely cost-free and autonomous.

🚀 The Mission

Students often use AI tools without knowing how their data is handled. ConsentIQ solves this by providing a "Traffic Light" risk system, allowing users to understand:

• Data Retention: How long is your data kept?
• Third-party Sharing: Is your data sold to partners?
• Biometric Usage: Does the app collect face/voice data?
• Targeted Advertising: Is your activity used for marketing?
• Data Deletion: Can you actually remove your data?

🛠️ Tech Stack (100% Local & Cost-Free)

• Frontend: React.js, Tailwind CSS, Framer Motion (Glassmorphism UI).
• Visuals: Three.js & WebGL (High-performance animated background).
• Backend: Node.js/Express (Secure Intelligence Proxy).
• AI Brain: Python (FastAPI), Transformers (HuggingFace DistilBERT).
• Stealth Scraper: Playwright-Stealth (Bypasses bot-detection on Google/OpenAI).
• Optimization: AMD ROCm support for local GPU acceleration.

✨ Key Features

🕵️ Stealth Intelligence Scanner

Bypasses restrictive "anti-bot" measures on major domains like OpenAI, Google Workspace, and Instagram. Uses human-like behavior simulation and randomized browser fingerprints.

🧠 Chunked NLP Brain

Analyzes large documents (20,000+ words) by extracting strategic chunks from the start, middle, and end. Uses local "Zero-Shot" classification to assign scores without any paid API calls.

📊 Intelligence Dashboard

Real-time analytical KPIs showing average safety scores, total apps analyzed, and interactive sorting for "High Risk" threats.

📄 Universal Extraction

Native support for HTML privacy pages and direct PDF downloads (corporate policies).

💻 Hardware Optimization (AMD ROCm)

ConsentIQ is designed to run locally on your hardware. If you have an AMD GPU, you can leverage ROCm to run the NLP models 10x faster than CPU.

See the ROCm_Guide.md for setup instructions.

⚙️ Installation & Usage

Prerequisites

• Node.js (v18+)
• Python 3.10+
• Chromium (for Playwright)

1. Setup the Repository

git clone https://github.com/yourusername/consentiq.git
cd consentiq
npm install

2. Setup the AI Brain (Python)

cd python-nlp
pip install -r requirements.txt
python -m playwright install chromium

3. Run Everything

Use the integrated development command to launch the dashboard and the AI Brain simultaneously:

npm run dev:all

The app will be live at http://localhost:3007.

📸 Screenshots

Dashboard Intelligence	Data Nutrition Label

---

Hackathon: AMD Slingshot Hackathon
Role: Privacy Specialist & Full-Stack Engineer
Cost: $0.00 (Fully Autonomous local inference)