SmishGuard: AI-Powered SMS Security Shield

SmishGuard is a serverless security backend that uses Artificial Intelligence to detect phishing, hate speech, and violent content in SMS messages. Built for the Microsoft Imagine Cup 2026.

Live Demo

API Endpoint: https://smishgaurd-app.azurewebsites.net/api/AnalyzeText (Note: Requires POST request with JSON body)

Architecture

• Cloud Provider: Microsoft Azure
• Compute: Azure Functions (Serverless Python 3.11)
• AI Engine: Azure AI Content Safety
• Development: VS Code + Azure Core Tools

How It Works

1. Ingest: The system receives an SMS text via HTTP POST.
2. Analyze: The text is sent to the Azure AI Content Safety neural network.
3. Detect: The AI scores the text for multiple risk categories:
   • Hate Speech
   • Violence
   • Self-Harm
   • Sexual Content
4. Respond: The API returns a JSON report with a "Suspicious" flag (True/False) and severity scores.

Tech Stack

• Python 3.11
• Azure Functions
• Azure Cognitive Services

Usage (Test with Thunder Client/Postman)

POST https://smishgaurd-app.azurewebsites.net/api/AnalyzeText

Body:

{
  "sms_text": "URGENT: Your bank account is locked. Click here: [www.fakebank.com](https://www.fakebank.com)"
}