Second Look

Inspiration

Online scams are becoming more convincing and more common, especially for older adults who may not be familiar with modern phishing tactics. As technology improves exponentially, scams are becoming even more difficult to detect. Data from the Canadian Anti-Fraud Centre shows the magnitude of the problem: Canadians lost $638 million to fraud in 2024, with reported losses since 2021 now surpassing $2 billion (StatsCan and Government of Canada).

Second Look was inspired by witnessing how easily scammers can trick people into clicking suspicious links, responding to phishing emails, or even entering sensitive information through fake bank websites. We wanted to create a tool that gives people a safe way to get a "second look" at messages or emails, providing users with peace of mind and helping them feel safe online.

What it does

Second Look is an interactive platform that helps users detect potential scams in messages by analyzing uploaded images of texts or emails. The app provides a risk assessment, displays a scam probability percentage score, and explains why a message might be unsafe, giving users actionable steps to continue.

Key features include:

• Scam Detector: Users upload a screenshot or image of a message or email to analyze for potential scams.
• Risk Dashboard: Displays overall scam risk and confidence level.
• Analysis Details: Breaks down red flags detected in the message, such as unverifiable sender information, unusually high pay offers, or generic content.
• Scam Type Identification: Classifies the type of scam detected, e.g., job offer scams or phishing attempts.
• Actionable Recommendations: Provides clear guidance on what to do next, including blocking the sender, verifying independently, and deleting suspicious messages.
• Visual Highlights: Suspicious content is clearly marked to help users understand why the message is risky.

How we built it

Second Look is powered by:

Frontend Framework & UI

• React (interactive and responsive user interface)
• Vite
• Chakra-UI (UI components & customizable styling)

AI & Machine Learning

• Google Gemini (analyzes text and images to detect potential scams)
• Optical Character Recognition (OCR) (extracts text from uploaded screenshots)

Backend & Data

• Node.js
• Temporary storage (images are stored securely for analysis and not retained)

Challenges we ran into

• Integrating image and text analysis through Gemini API and integrating it into the backend
• Adding the chatbot functionality. We wanted to provide explanations and guidance in real-time. Integrating and making the chatbot respond accurately to varied messages. Giving the chatbot memory and context.

Accomplishments that we're proud of

• First and second hackathons for both members!
• Building a fully functional platform in 24 hours that analyzes screenshots and emails, provides risk scores, highlights red flags, and explains why a message might be unsafe

What we learned

• How to integrate and use Gemini AI in a project
• Using React to build out a responsive and interactive front end
• How to create a Gemini chatbot to answer relevant questions related to the screenshots / images

What's next for Second Look

• Email account integration to automatically scan emails, label high potential scams, and move them to a separate folder for manual review.
• Educational dashboard to track common scams and teach users how to spot them for future protection.
• User accounts & database: Allow users to sign in, save past uploads and chatbot conversations, and continue without starting over each time.

Built With

• chakra
• gemini
• node.js
• react
• vite