Inspiration
Phishing and online scams are increasing rapidly, and many users—especially students and non-technical users—fall victim because malicious messages often look normal at first glance. I wanted to build something that acts like a real-time security assistant and helps users instantly understand whether a message is safe or dangerous.
What it does
NeuroShield AI is a real-time phishing and scam detection system powered by Google Gemini AI. Users can paste any message, email, or suspicious text, and the system analyzes it to generate a structured cybersecurity report including risk score, threat type, red flags, and safety advice.
How I built it
I built the frontend using React, TypeScript, and Tailwind CSS on the Lovable AI platform. For intelligence, I integrated Google Gemini API to analyze text and return structured security insights. The UI is designed as a cybersecurity dashboard with real-time analysis experience.
Challenges I ran into
The main challenge was ensuring consistent AI responses in a structured format (risk score, threat type, explanation). I solved this by carefully designing the prompt sent to Gemini and refining response parsing.
What I learned
I learned how to integrate large language models into real applications, how prompt engineering affects output quality, and how to design user-friendly cybersecurity tools that feel like real-world SaaS products.
What's next
I plan to improve detection accuracy, add browser/email extensions, and expand it into a full cybersecurity assistant platform for real-time protection.
Built With
- ai
- google-gemini-api
- lovable
- react
- tailwind-css
- typescript
- vite
Log in or sign up for Devpost to join the conversation.