-
Detect · Explain · Protect
Inspiration
The inspiration behind PhishGuard AI came from the growing number of phishing and scam messages people encounter daily, especially through SMS, email, and messaging platforms. Many of these scams rely on urgency and deception, making them difficult for everyday users to identify. We realized that while cybersecurity tools exist, most are either too technical or not easily accessible. We wanted to build a solution that simplifies cybersecurity — something that doesn’t just detect threats, but also helps users understand them in real time. This led us to create PhishGuard AI: a tool that empowers users to recognize and avoid scams using simple, intelligent analysis.
Problem
Phishing and scam messages are one of the most common cybersecurity threats today, especially on messaging platforms like WhatsApp, SMS, and email. Many users: 1.Cannot easily identify scam messages 2.Fall for urgency tactics (“act now”, “verify immediately”) 3.Click malicious links unknowingly Existing tools are often too technical or inaccessible to everyday users. PhishGuard AI solve this problem.
What it does(Solution)
PhishGuard AI is a simple, browser-based tool that helps users instantly detect and understand phishing attempts. Users can paste any message, email, or link, and the system: .Analyzes it using AI .Assigns a risk score (0–100) .Explains why it may be dangerous .Highlights suspicious phrases and links It also goes further by detecting suspicious URLs & rewriting messages into safer versions.
Key Features
🔐 AI-Powered Risk Analysis: generates a phishing risk score with clear explanations.
⚠️ Suspicious Phrase Detection: highlights risky patterns like urgency (“urgent”, “act now”) and sensitive requests (“verify password”)
🔗 Smart Link Scanner: detects potentially malicious links such as shortened URLs (bit.ly, tinyurl) and suspicious domains.
🧠 AI Explanation Engine: explains why a message is dangerous in simple language.
✍️ Safe Rewrite Feature: transforms suspicious messages into safe, non-malicious versions.
🎨 Modern Interactive UI: Animated results, Visual risk meter and Clean and intuitive design.
User pastes a message, the app sends it to an AI model and AI returns: 1.risk score
- explanation 3.suspicious phrases The UI displays results visually. Optional: user rewrites message safely
How we built it
We built PhishGuard AI as a lightweight, browser-based application using a simple but effective tech stack. 1.The frontend was developed using HTML and CSS, focusing on a clean and intuitive user interface. Vanilla JavaScript handled user interactions, input processing, and dynamic updates. 2.For the core intelligence, we integrated an AI API to analyze user-submitted messages. The AI evaluates the text, assigns a phishing risk score, identifies suspicious phrases, and generates a clear explanation of potential threats. 3.We also implemented additional logic to detect suspicious links and highlight risky patterns commonly used in scams. 4.To ensure reliability during demos, we included a fallback analysis system that simulates results if the API is unavailable.
- Finally, we enhanced the user experience with animations, a visual risk meter, and a “safe rewrite” feature that transforms suspicious messages into secure alternatives.
Challenges we ran into
1.Handling inconsistent AI responses (JSON parsing issues) 2.Ensuring the app still works without API (fallback logic) 3.Designing a clean UI within limited time
Accomplishments that we're proud of
1.Built a fully functional AI-powered security tool in under 24 hours 2.Created a polished and interactive user experience 3.Added multiple layers of analysis (AI + heuristic detection)
Impact
1.Accessible to non-technical users 2.Educational by explaining threats 3.Preventive by catching scams early This is especially valuable in regions where digital scams are increasing rapidly.
What we learned
1.Throughout this project, we learned how to balance simplicity with functionality — building a tool that feels powerful while remaining easy to use. 2.We gained hands-on experience integrating AI into a frontend application and handling real-world challenges such as inconsistent API responses and error handling. 3.We also learned the importance of user experience in cybersecurity tools. Clear visual feedback, simple explanations, and intuitive design can make complex concepts accessible to a wider audience. 4.Additionally, we realized that in a hackathon setting, presentation and reliability are just as important as technical complexity. Ensuring the app works smoothly, even under failure conditions, was a key takeaway. 5.Overall, this project strengthened our ability to rapidly prototype impactful solutions using modern web technologies and AI.
What's next for PhishGuard AI — Real-Time Scam & Phishing Detector
1.Real-time browser extension 2.Integration with messaging platforms 3.More advanced phishing detection models 4.Multi-language support
Log in or sign up for Devpost to join the conversation.