Machine Learning Phishing Email Detection Tool

Machine Learning-Powered Phishing Detection for Microsoft Outlook add-in

Get Started

Email Analysis Tool

A Microsoft Outlook Add-in tool that allows user to perform user sanitisation check on emails

The Email Analysis Tool is designed to integrate seamlessly with Microsoft Outlook via a custom add-in. The tool allows users to analyze suspicious emails with a single click before interacting with or reporting them as phishing. The tool extracts key details such as the sender domain, subject, body content, embedded URLs, and email authentication headers.

These details are sent to a backend system that performs a three-layered analysis — Content, Link, and Header Analysis . The backend uses a hybrid machine learning model powered by BERT, combined with reputation checks from VirusTotal and WHOIS, and authentication validation using SPF, DKIM, and DMARC headers.

In addition to the primary detection model, the tool offers an alternative analysis option using the Perplexity AI API, providing a second opinion powered by advanced language understanding along with a concise explanation of the email classification.

The results are displayed directly in the Outlook task pane for real-time analysis, giving users clear insights into the email's classification, link safety, and authentication status. This enables users to make informed decisions and quickly determine if an email poses a potential threat.

Interface on Outlook

Key Features

What makes the Email Analysis Tool powerful and effective

Seamless Outlook Integration

Launches directly within Microsoft Outlook as a task pane add-in, allowing users to analyze emails with a single click without leaving their inbox.

Intelligent Content Analysis

Uses a hybrid machine learning model powered by BERT to understand the email's language and detect phishing or spam patterns in the subject and body.

Advanced Link Analysis

Inspects embedded URLs using VirusTotal for threat detection and WHOIS for domain age to assess the safety of links and flag suspicious domains.

Header Authentication Checks

Validates SPF, DKIM, and DMARC results to ensure the email is truly from its claimed sender and not spoofed.

Perplexity AI Integration

Offers an optional second-opinion analysis powered by the Perplexity API, including a short explanation to help users understand the classification decision.

Real-Time Results

Displays the complete classification, link safety insights, and authentication summary instantly within the Outlook task pane interface.

Documentation

Browse the full reports and documentation related to the development of the Email Analysis Tool

Research Report

Outlines the research done during the early stages of the project.



View Research Report

Project Specification

Details the system design, architecture, tools used, and planned development milestones for the phishing detection tool.

View Documentation

Final Report

Comprehensive documentation of the final system, implementation, testing, and limitations of the Email Analysis Tool.

View Final Report

Showcase

Explore real examples of how the Email Analysis Tool evaluates emails inside Outlook.

Legitimate Email Example

Legitimate Email Example

Safe emails are displayed to the user with a green "Safe Email" label and confidence score.

Spam Detection

Spam Detection

Demonstrates how the system identifies and flags potential spam messages.

Classification

Classification

Shows the ML model's classification results with confidence scores.

Email Authentication

Email Authentication

Displays SPF, DKIM, and DMARC results to verify sender authenticity.

Link Risk Analysis

Link Analysis

Evaluates URLs in the email using VirusTotal and WHOIS data for comprehensive risk assessment.

Perplexity Analysis

Perplexity Analysis

Provides AI-powered second opinion with natural language explanation.

Let's Connect

Feel free to reach out through any of these platforms for collaboration, questions, or just to say hello!