AkifSivas58/SuperSecret

For Huckathon at Hacettepe University

0

stars

62

commits

JavaScript

primary language

Nov 27, 2025

updated

README

🛡️ SuperSecret - AI-Guarded Ephemeral Messaging

Python AI Security Realtime

SuperSecret is a secure, real-time messaging platform developed during the HUCKATHON. It features ephemeral chats that self-destruct upon closure and an AI-powered content moderation system that filters toxic language, spam, and malicious URLs in real-time.


📸 Demo Video

https://github.com/user-attachments/assets/25ca805f-ace2-4b4e-afdb-71148f7a43cc

Demo Highlights:

  • 0:46 - Real-time "Mental Connection" request & acceptance.
  • 1:06 - Trying to connect to a busy user triggers the "Mind Occupied" blocking protocol.
  • 1:20 - Multi-session concurrency (4 simultaneous users).

🚀 Key Features

🔒 Privacy & Security by Design

  • Ephemeral Messaging: Chats are stored in temporary, dynamic SQLite tables (Chat_{id}). Once the session ends, the table is DROPPED, ensuring zero data retention.
  • End-to-End Encryption: All messages are encrypted using Fernet (Symmetric Encryption) before hitting the database.
  • Secure Authentication: User passwords are hashed using Bcrypt.

🧠 "Mental Connection" Protocol (Session Locking)

  • Exclusive Connectivity: Following the project's lore, when two users establish a "Mental Connection" (start a chat), their status instantly updates to Busy.
  • Request Blocking: The system enforces strict session isolation; any incoming requests to a user already in a mental connection are automatically rejected, ensuring an uninterrupted and private experience.

🤖 AI-Powered Content Guard (MessageChecker.py)

Unlike standard chat apps, SuperSecret analyzes every message before it is delivered using NLP models:

  1. Translation Layer: Automatically translates TR messages to EN using Helsinki-NLP for accurate analysis.
  2. Toxicity Detection: Blocks offensive content using RoBERTa.
  3. Phishing/Malicious URL Shield: Detects harmful links using BERT.
  4. Spam Filter: Identifies and blocks spam patterns.

⚡ Modern Real-Time UI

  • Socket.IO: Instant message delivery and live status updates (Online/Busy/Offline).
  • Cyberpunk UI: A responsive, dark-mode interface with particle network animations (HTML5 Canvas).

🛠️ Tech Stack

Backend

  • Framework: Python Flask
  • Real-time Engine: Flask-SocketIO
  • Database: SQLite (Dynamic Table Architecture)
  • Security: Cryptography (Fernet), Bcrypt, JWT (JSON Web Tokens)

Artificial Intelligence

  • Libraries: transformers, torch
  • Models:
    • kmack/malicious-url-detection
    • mshenoda/roberta-spam
    • cardiffnlp/twitter-roberta-base-offensive
    • Helsinki-NLP/opus-mt-tc-big-tr-en

Frontend

  • Core: HTML5, CSS3 (Custom Variables), JavaScript (ES6+)
  • Styling: Bootstrap 5, FontAwesome [cite: 4]
  • Visuals: Custom Particle Network Animation (script.js)

⚙️ Installation & Setup

  1. Clone the repository
    git clone [https://github.com/yourusername/SuperSecret.git](https://github.com/yourusername/SuperSecret.git)
    cd SuperSecret
    
  2. Create a Virtual Environment
    python -m venv venv
    source venv/bin/activate  # On Windows: venv\Scripts\activate
    
  3. Install Dependencies
    pip install -r requirements.txt
    
  4. Configure Environment Create a .env file in the root directory:
    SECRET_KEY=your_secret_key
    JWT_SECRET_KEY=your_jwt_secret
    FERNET_KEY=your_fernet_key
    
  5. Run the Application
    python app.py
    

The server will start at http://localhost:5000

Contributors

Erdem93

22 commits

AkifSivas58

16 commits

davutsimsekk

12 commits

b2230356003

8 commits

AkifSivas58/SuperSecret

For Huckathon at Hacettepe University

0

stars

62

commits

JavaScript

primary language

Nov 27, 2025

updated

README

🛡️ SuperSecret - AI-Guarded Ephemeral Messaging

Python AI Security Realtime

SuperSecret is a secure, real-time messaging platform developed during the HUCKATHON. It features ephemeral chats that self-destruct upon closure and an AI-powered content moderation system that filters toxic language, spam, and malicious URLs in real-time.


📸 Demo Video

https://github.com/user-attachments/assets/25ca805f-ace2-4b4e-afdb-71148f7a43cc

Demo Highlights:

  • 0:46 - Real-time "Mental Connection" request & acceptance.
  • 1:06 - Trying to connect to a busy user triggers the "Mind Occupied" blocking protocol.
  • 1:20 - Multi-session concurrency (4 simultaneous users).

🚀 Key Features

🔒 Privacy & Security by Design

  • Ephemeral Messaging: Chats are stored in temporary, dynamic SQLite tables (Chat_{id}). Once the session ends, the table is DROPPED, ensuring zero data retention.
  • End-to-End Encryption: All messages are encrypted using Fernet (Symmetric Encryption) before hitting the database.
  • Secure Authentication: User passwords are hashed using Bcrypt.

🧠 "Mental Connection" Protocol (Session Locking)

  • Exclusive Connectivity: Following the project's lore, when two users establish a "Mental Connection" (start a chat), their status instantly updates to Busy.
  • Request Blocking: The system enforces strict session isolation; any incoming requests to a user already in a mental connection are automatically rejected, ensuring an uninterrupted and private experience.

🤖 AI-Powered Content Guard (MessageChecker.py)

Unlike standard chat apps, SuperSecret analyzes every message before it is delivered using NLP models:

  1. Translation Layer: Automatically translates TR messages to EN using Helsinki-NLP for accurate analysis.
  2. Toxicity Detection: Blocks offensive content using RoBERTa.
  3. Phishing/Malicious URL Shield: Detects harmful links using BERT.
  4. Spam Filter: Identifies and blocks spam patterns.

⚡ Modern Real-Time UI

  • Socket.IO: Instant message delivery and live status updates (Online/Busy/Offline).
  • Cyberpunk UI: A responsive, dark-mode interface with particle network animations (HTML5 Canvas).

🛠️ Tech Stack

Backend

  • Framework: Python Flask
  • Real-time Engine: Flask-SocketIO
  • Database: SQLite (Dynamic Table Architecture)
  • Security: Cryptography (Fernet), Bcrypt, JWT (JSON Web Tokens)

Artificial Intelligence

  • Libraries: transformers, torch
  • Models:
    • kmack/malicious-url-detection
    • mshenoda/roberta-spam
    • cardiffnlp/twitter-roberta-base-offensive
    • Helsinki-NLP/opus-mt-tc-big-tr-en

Frontend

  • Core: HTML5, CSS3 (Custom Variables), JavaScript (ES6+)
  • Styling: Bootstrap 5, FontAwesome [cite: 4]
  • Visuals: Custom Particle Network Animation (script.js)

⚙️ Installation & Setup

  1. Clone the repository
    git clone [https://github.com/yourusername/SuperSecret.git](https://github.com/yourusername/SuperSecret.git)
    cd SuperSecret
    
  2. Create a Virtual Environment
    python -m venv venv
    source venv/bin/activate  # On Windows: venv\Scripts\activate
    
  3. Install Dependencies
    pip install -r requirements.txt
    
  4. Configure Environment Create a .env file in the root directory:
    SECRET_KEY=your_secret_key
    JWT_SECRET_KEY=your_jwt_secret
    FERNET_KEY=your_fernet_key
    
  5. Run the Application
    python app.py
    

The server will start at http://localhost:5000

Contributors

Erdem93

22 commits

AkifSivas58

16 commits

davutsimsekk

12 commits

b2230356003

8 commits

Languages

JavaScript

40.8%

Python

29.5%

CSS

22.1%

HTML

7.5%