It provides :

• Fraud Detection Services (Sentinel)
• Financial Guidance (Moneto)
• Portfolio Management (Future Project ---> Vigilant)

Next-Generation Financial Intelligence Platform
🌐 Live Demo: https://financewise-ldv2.onrender.com/
📄 Presentation: View Slides

FinanceWise is an AI-powered financial platform combining two robust systems:

1. Sentinel 🛡️ - Advanced Fraud Detection System
2. Moneto 💬 - Intelligent Financial Advisor

{
  "Real-time transaction analysis",
  "Deep learning model with 99.9% accuracy",
  "Suspicious pattern recognition",
  "AI-generated risk explanations"
}

Issues

• Trained on Dataset with a 1.29% Fraud Rate, thus resulting in some anomalies which will be rectified in a few updates.

{
  "Llama-3 70B-powered financial guidance",
  "Real-time market insights",
  "Personalized wealth management",
  "Risk assessment strategies",
}

{

  "*** COMING SOON ***"

}

Backend

Frontend

APIs & Services

Personal Model
SENTINAL : Self-Created and Self Trained

• Python 3.9+
• pip package manager
• Git version control

1. Clone repository:

   git clone https://github.com/Jayesh-Dev21/FinanceWise.git
   cd FinanceWise

2. Create virtual environment:

   python -m venv venv
   # Linux/MacOS
   source venv/bin/activate
   # Windows
   venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Configure environment variables: Create a .env file in the root directory:

   GROQ_API_KEY="your_groq_api_key_here"
   FLASK_DEBUG=0
   PORT=1200

5. Download ML models:

   # Place scaler.pkl and fraud_detection_model.pth in root directory

python app.py

Open your browser and navigate to:

• Home Page: http://localhost:1200
• Sentinel Dashboard: http://localhost:1200/fraud_detection
• Moneto Interface: http://localhost:1200/financial_advice

{
  "amount": 1500,
  "credit_limit": 5000,
  "use_chip": "Chip Transaction",
  "merchant_category": "Online Retail",
  "transaction_date": "2024-03-15"
}

{
  "message": "Should I invest 20% of my salary in cryptocurrency?"
}

{
  "probability": 0.92,
  "prediction": "Fraud",
  "explanation": "Risk assessment: 92.3% suspicious activity likelihood"
}

{
  "response": "While cryptocurrency can offer high returns, I recommend..."
}

FinanceWise/
├── public/                    # Frontend assets
│   ├── static/                # Static resources
│   │   ├── css/               # Stylesheets
│   │   ├── js/                # Client-side scripts
│   │   └── images/            # Visual assets
│   └── *.html                 # Application views
├── app.py                     # Flask application
├── requirements.txt           # Dependency list
├── scaler.pkl                 # Feature scaler
├── fraud_detection_model.pth  # Trained model
└── .env                       # Environment configuration

• Harsh Gupta – Sentinel AI Model Training & Optimization
• Jayesh Puri – Moneto API Integration & LLM Configuration in Sentinel
• Jaydeep Pokahriya – Full Web Stack Development (Flask Backend + Frontend) and GitHub Handling
• Krrish – Project Presentation

• Groq: For ultra-low latency LLM inference
• Hugging Face: For transformer models (pre-trained models)
• Render: For reliable cloud hosting
• PyTorch Community: For deep learning framework

• Obtain a Groq API key from Groq Cloud.
• Do not commit the .env file to version control.
• For production deployment, set FLASK_DEBUG=0.