Explainable AI • Email Security

Detect phishing with evidence, not guesses.

PHISHLENS classifies emails as phishing/spam or legitimate and explains the decision using evidence tokens, reason categories, and highlighted suspicious text.

Evidence tokensReason categoriesHighlightsMetrics

Problem

Phishing looks “normal”

Attackers mimic trusted brands and use urgency + links to trick users into clicking or sharing credentials.

Solution

ML classification

PHISHLENS learns patterns from real datasets to output a phishing probability for each email.

Uniqueness

Explainability first

Instead of only a label, PHISHLENS shows evidence, categories, and highlighted text to support trust.

How it works

Input emailCleaningVectorizer (TF-IDF)ClassifierExplainability output

The website sends your email subject/body to the Python backend. The model returns a probability plus explainability data that the UI turns into an evidence dashboard.