ScamShield AI powered by Gemma

*Build a complete Hacktoberfest Open-Source AI project named ScamShield AI.

IMPORTANT: Google Gemma must be the actual core AI model used in this application.

Model: google/gemma-4-E2B-it.

Project concept:
An Indian e-commerce shopping platform inspired by Meesho, with an integrated AI scam detection system.

Build the following:

1. Create an attractive shopping website with product categories, product cards, product descriptions, seller profiles and customer reviews.

2. Integrate the actual Gemma model using Hugging Face Transformers or a supported inference provider.

3. Create an AI scam analyser where users can submit:

• Product descriptions.
• Seller information.
• WhatsApp messages.
• SMS messages.
• Suspicious payment requests.
• Shopping offers.

1. Make Gemma analyse the submitted content and identify suspicious patterns.

2. Generate structured results:

• Risk score from 0 to 100.
• Low, Medium or High risk.
• Suspicious indicators.
• Detailed AI explanation.
• Safety recommendations.

1. Add a scam detection dashboard with scan history and charts.

2. Use Python and Flask for the backend, HTML, CSS and JavaScript for the frontend, and SQLite for storage.

3. Do not use paid OpenAI APIs.

4. Do not generate fake AI responses or use only predefined if-else conditions. The actual Gemma model must process user input.

5. Combine Gemma's semantic analysis with transparent fraud detection rules. Clearly distinguish model-generated assessment from rule-based signals.

6. Create a Gradio-based demo version suitable for Hugging Face Spaces.

7. Provide the complete GitHub project structure, source code, requirements.txt, README.md and installation instructions.

8. Include sample genuine and suspicious messages for testing.

9. Ensure all features work and provide instructions for deployment.

10. Prepare the final Hacktoberfest submission using the official Week 4 template.

Build step by step. Do not claim that the model or demo is working until it has been successfully tested.This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine*

What I Built

Demo

Code

How I Built It

Why Does Open Innovation Matter?

My Agent Session

The Spirit of It

Prize Categories