Inspiration
Inspired by the Trojan Horse from Greek mythology, Trojan protects grandparents from digital scams disguised as innocent messages. Our goal is simple: let Grandma open every gift safely.
What it does
Trojan is an AI-powered scam guard that analyzes suspicious texts, emails, and screenshots, classifying them as SAFE, SUSPICIOUS, or SCAM. It explains risks through a calm voice assistant and instantly alerts family members through iMessage.
How we built it
We built Trojan using Python, FastAPI, Google Gemini, MongoDB Atlas, and macOS automation. It combines voice activation, message and email reading, screenshot analysis, and real-time family notifications into one system.
Challenges we ran into
Integrating multiple macOS services, extracting message content, handling voice recognition, and maintaining reliable communication between devices within a tight hackathon deadline.
Accomplishments that we're proud of
Building a unified AI security assistant that works across Mac and iPhone, provides understandable scam explanations, and keeps family members informed without interacting with potentially dangerous content.
What we learned
We learned how to integrate AI with real-world applications, work with macOS automation, build APIs, and prioritize functionality while collaborating under time pressure.
What's next for Trojan
Improve voice quality, introduce a polished interactive interface, build a real-time family dashboard, and expand device support to make digital safety accessible to everyone.
