Phishy Phighters

Phishy Phighters Homepage
Phishy Phighters Interactive Gemini AI Learning Module
Phishy Waters Extension

Inspiration

We noticed how often misinformation and phishing attempts sneak into people’s daily browsing, especially on social media. Many users don’t fact-check or know what red flags to look for. We wanted to make cybersecurity practices accessible, engaging, and interactive, so anyone can learn while they browse. That’s how Phishy Waters (the web extension) and Phishy Phighters (the web app) were born.

What it does

Phishy Waters (browser extension):
Scrapes page text and links, filters them, and uses Fact Check API to flag suspicious or false content. Currently, it works on Reddit, having the capability to highlight misleading claims and dangerous links in real time.
Phishy Phighters (web app):
An interactive classroom reimagined. Instead of static quizzes, a Gemini-powered chatbot (“Phishy”) guides users through phishing basics, misinformation spotting, and safe browsing by asking questions, explaining mistakes, and adapting to the user’s level.

Together, they create a hands-on safety experience: learn, test, and defend yourself in real time.

How we built it

Extension:
- Built a page content + link scraper.
- Passed page text into Google’s Fact Check API to cross-reference against misinformation databases.
- ~~Displayed results in a browser popup for instant user feedback.~~ (Disclaimer: we were not able to get this to work so the results can be displayed in the inspect of the popup)
Web app:
- Next.js React frontend with Tailwind for styling.
- Deployed on Vercel, and utilized Vercel Typescript Template to start!
- Integrated Google Generative AI (Gemini) API to power the chatbot.
- Managed message history and context for smooth multi-turn conversations.
- Designed mascot “Phishy the Fish” to make cybersecurity less intimidating.

Challenges we ran into

Getting the Google Fact Check API to reliably parse scraped text (needed filtering + formatting).
Extension works only on Reddit, since other sites render DOM differently.
Integrating Gemini required careful handling of conversation state.
Balancing educational accuracy with fun, engaging design.
Highlighting the specific text on the page that is false.

Accomplishments that we're proud of

Building a working prototype of both a browser extension and a chatbot app in < 24 hrs.
Successfully connecting Google APIs (Fact Check + Gemini).
Creating a fun and accessible learning experience that makes security approachable.

What we learned

How to structure a browser extension with background scripts and popup UIs.
Effective use of Google Fact Check API.
Best practices for maintaining conversation context in LLM-powered chatbots.

What's next for Phishy Phighters

Expand extension support to Twitter/X, Facebook, YouTube.
Add UI spoof detection and more phishing indicators.
Improve chatbot with adaptive learning paths and personalized feedback.
Combining extension + chatbot for one seamless experience.
Release as a free public tool for safer browsing.

For Base44 Challenge 2

Misinformation and phishing are widespread online, especially on social media, where people often scroll quickly without verifying sources or recognizing red flags. This creates an environment where false claims spread unchecked and malicious links trick users into giving away sensitive information. Many existing security tools are either too technical or too passive, leaving everyday users without engaging, accessible guidance.

Our solution combines a browser extension (Phishy Waters) with a companion web app (Phishy Phighters) to deliver cybersecurity education in real time. The extension scrapes page text and links, filters them through Google’s Fact Check API, and highlights misleading or dangerous content directly on social media sites (currently only works on Reddit and news sites!). The web app transforms learning into an interactive experience, where our Gemini-powered chatbot (“Phishy”) teaches phishing awareness and misinformation detection through adaptive Q&A. Together, they provide both practical protection while browsing and an engaging learning platform. We aim to make cybersecurity education scalable, approachable, and impactful for millions of users worldwide by expanding to all social media sites and adding features like spoof detection and personalized learning paths.