🛡️ PHISH-SENTRY v2.4

Fully Client-Side Phishing Email Detector
Powered by Scikit-learn + Pyodide WebAssembly

Overview

PHISH-SENTRY is a fully client-side Machine Learning application that detects phishing emails directly in your browser. It uses Pyodide (Python compiled to WebAssembly) to run a Scikit-learn ML pipeline — no backend, no API keys, no data ever leaves your browser.

Run Locally

Prerequisites: Node.js (v18+)

npm install
npm run dev

The app will start at http://localhost:3000.

Architecture

• Frontend: React 19 + Vite + Tailwind CSS v4
• ML Engine: Pyodide (Python 3.12 WASM) + Scikit-learn + Pandas
• Model: TF-IDF Vectorizer → Multinomial Naive Bayes Pipeline
• Privacy: Zero data exfiltration — all inference runs locally in-browser

No API Keys Required

This application runs entirely client-side. No API keys, no backend servers, no external AI service calls. The only network requests are:

• Fetching the Pyodide WASM runtime from the official CDN (one-time)
• Downloading scikit-learn/pandas wheels from the Pyodide package index (one-time)