AI Email Safety Scanner (Phishing Detection RAG System)

A Python-based system that scans your Gmail inbox for phishing emails using a combination of machine learning, retrieval-augmented generation (RAG), and an LLM. Results are displayed in a Streamlit dashboard.

---

Features

• Real-time phishing email detection
• Gmail inbox monitoring
• ML-based classification
• Retrieval-Augmented Generation (RAG)
• AI-generated threat explanations
• Streamlit dashboard visualization

---

How it works

Emails are fetched from Gmail via the API, cleaned and preprocessed, then passed through a trained ML classifier. Suspicious emails are further analyzed using a FAISS-based RAG retrieval step and an LLM (OpenAI or Anthropic) to generate a human-readable threat explanation. Alerts can be sent automatically, and everything is viewable in the dashboard.

Gmail API → Preprocessing → ML Detection → FAISS Retrieval → LLM Analysis → Dashboard

---

Tech Stack

• Python
• Scikit-learn
• FAISS
• LangChain
• OpenAI / Anthropic APIs
• Streamlit
• Gmail API

---

Project structure

├── app.py               # Streamlit dashboard
├── agent.py             # LangChain pipeline orchestration
├── gmail_fetcher.py     # Fetches emails via Gmail API
├── preprocess.py        # Cleans and extracts email text
├── detector.py          # ML phishing classifier
├── rag.py               # FAISS-based RAG retrieval
├── llm.py               # LLM integration
├── notify.py            # Sends alerts for flagged emails
├── output.py            # Formats results
├── auto_scan.py         # One-time inbox scan
├── watch_inbox.py       # Continuous inbox monitoring
├── auth_fix.py          # OAuth helper
├── phishing_model.pkl   # Pre-trained scikit-learn model
├── rag_index.faiss      # FAISS vector index
├── rag_vectorizer.pkl   # Vectorizer for RAG
├── rag_examples.json    # Example phishing cases
├── requirements.txt
├── env.example
└── SETUP.md             # Gmail API setup guide

---

Setup

1. Clone and install

git clone https://github.com/akhilandeshwari2006/Phishing-Detection-RAG-System.git
cd Phishing-Detection-RAG-System

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

pip install -r requirements.txt

2. Environment variables

cp env.example .env

Add your API keys to .env:

OPENAI_API_KEY=...
ANTHROPIC_API_KEY=...

3. Gmail API credentials

See SETUP.md for the full walkthrough. In short:

• Create a Google Cloud project and enable the Gmail API
• Set up an OAuth 2.0 Desktop App credential
• Download and save it as credentials.json in the project root

4. Authenticate

python gmail_fetcher.py

A browser window will open for Gmail login. After that, token.json is saved and you won't need to log in again.

---

Usage

Command	What it does
python gmail_fetcher.py	Fetch emails and save to emails.json
python auto_scan.py	One-time scan of your inbox
python watch_inbox.py	Continuously monitor for new emails
streamlit run app.py	Open the dashboard at localhost:8501

---

Screenshots

Phishing Email Detection

Safe Email Detection

Notes

• credentials.json and token.json are gitignored — never commit them
• The app works with either OpenAI or Anthropic; set whichever key you have in .env