AWS Builder Center
Building IsItLegit: Protecting Everyday Users from Phishing and Scams Using AWS

Building IsItLegit: Protecting Everyday Users from Phishing and Scams Using AWS

A look into how I built a privacy-first cybersecurity web app that traps zero-day phishing evasion tricks and verifies genuine bank 2FA alerts using client-side heuristics, AWS Amplify, and Amazon Bedrock.

Student
Every day, millions of people receive urgent SMS notifications, suspicious parcel tracking links, and unexpected bank OTP codes. Standard browser blacklists often fail against zero-day phishing tricks like disguised IP addresses, spoofed credentials, and invisible Unicode characters. Even worse, typical security tools treat genuine bank transaction alerts the same as malware, causing unnecessary panic.
To solve this, I built IsItLegit—an open-source cybersecurity intelligence engine and Progressive Web App designed to let anyone inspect suspicious links, emails, and alerts without ever clicking them.

How I Built It

As a solo developer, I wanted an architecture that combined robust client-side heuristics with cloud AI reasoning while guaranteeing 100% user privacy.
  1. Client-Side Evasion Defense: IsItLegit runs in-browser algorithms to de-obfuscate DWORD decimal IPs, detect homoglyph brand impersonation, calculate Shannon entropy on domains, and verify genuine 2FA OTP codes.
  1. AI Reasoning with Amazon Bedrock: For complex, deceptive social engineering lures, I connected the backend to Amazon Bedrock using the  @aws-sdk/client-bedrock-runtime . Leveraging Anthropic Claude 3.5 Sonnet allows the app to deconstruct subtle coercion tactics and explain the threat in plain English.
  1. Serverless Delivery: I used the open-source AWS SAM CLI to define the Lambda function and API Gateway in a clean SAM template, using its native esbuild support to bundle modern TypeScript.
  1. Deploying with AWS Amplify: I configured AWS Amplify Hosting using an  amplify.yml  build specification. Amplify made it effortless to automate the build pipeline and inject critical HTTP security headers like  Strict-Transport-Security  and  X-Frame-Options: DENY  to protect users against clickjacking.

Key Takeaway

Building during the hackathon showed me how powerful serverless AI workflows can be for cybersecurity. Combining automated client-side traps with Amazon Bedrock and AWS Amplify allowed me to ship a fast, zero-storage tool that gives everyday users peace of mind before they click.
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article