TruthLens — Multi-Layer AI Fake News & Spam Detection System

What It Does

TruthLens detects fake news and spam using three independent AI layers working together.

Layer	Technology	Accuracy
ML Layer	TF-IDF + Logistic Regression	99.3%
Wikipedia Layer	Cosine Similarity	Live API
AI Layer	Groq LLaMA 3.3-70B	Explanation

Features

• 99.3% fake news detection accuracy
• 98.1% spam detection accuracy
• Live Wikipedia cross-verification
• Groq LLaMA 3.3-70B AI explanations
• Confidence gauge meter
• Credibility score breakdown
• Real-time suspicious word detection
• URL article fetcher and analyzer
• Batch headline analyzer
• User authentication login and register
• Search history with timestamps
• Usage statistics dashboard
• Download analysis report
• WhatsApp share button
• GDPR compliant data deletion
• Source credibility checker
• Language detection
• Dark and light mode toggle
• Mobile responsive design

Tech Stack

• Backend: Python Flask Scikit-learn SQLite
• ML: TF-IDF Vectorizer Logistic Regression
• APIs: Wikipedia REST API Groq Cloud API
• Frontend: HTML5 CSS3 JavaScript Chart.js

How To Run

# 1. Clone the repository
git clone https://github.com/Janani7975/TruthLens.git
cd TruthLens

# 2. Create virtual environment
python -m venv venv
venv\Scripts\activate

# 3. Install dependencies
pip install -r requirements.txt

# 4. Create .env file with your keys
# GROQ_API_KEY=your_groq_key_here
# SECRET_KEY=truthlens2026

# 5. Download datasets from Kaggle into data/ folder
# Fake and Real News Dataset → data/Fake.csv and data/True.csv
# SMS Spam Collection → data/spam.csv

# 6. Train the models one time only
python train_models.py

# 7. Run the application
python main.py

Open browser at http://127.0.0.1:5000

Demo Login

• Email: demo@truthlens.com
• Password: demo123

Project Structure

TruthLens/ ├── app/ │ ├── routes.py │ └── services/ │ ├── ml_service.py │ ├── wikipedia_service.py │ ├── ai_report_service.py │ ├── combination_engine.py │ ├── url_service.py │ ├── source_service.py │ └── language_service.py ├── database/ │ └── db.py ├── static/ │ ├── style.css │ └── script.js ├── templates/ │ ├── index.html │ └── login.html ├── train_models.py ├── main.py └── requirements.txt

API Endpoints

Method	Endpoint	Description
POST	/api/analyze-full	Full 3-layer analysis
POST	/api/detect/fake-news	ML only detection
POST	/api/detect/spam	Spam detection
POST	/api/analyze-url	URL article analyzer
POST	/api/analyze-batch	Batch analyzer
GET	/api/history	Search history
GET	/api/statistics	Usage stats
DELETE	/api/privacy/delete-data	GDPR delete

Model Performance

Metric	Fake News	Spam
Accuracy	99.3%	98.1%
Precision	97.1%	97.8%
Recall	97.3%	94.2%
F1 Score	97.2%	96.0%