AI Scam Detector

• 

  "Main Dashboard: The user-friendly interface where users can paste messages or upload screenshots for instant AI analysis."

• 

  "Under the Hood: A glimpse of the Python code integrating Google Gemini API logic with Streamlit framework."

What it does## 💡 Inspiration

In today's digital world, online scams are increasing every day. Many innocent people in my community receive fake lucky draw messages, fraudulent job offers, and manipulated payment screenshots. I wanted to build a solution that empowers common people to verify these messages instantly using AI, without needing technical knowledge.

🤖 What it does

AI Scam Detector is an intelligent web application that acts as a personal digital guard.

• Text Analysis: Users can paste any suspicious message (SMS, WhatsApp, Email), and the AI analyzes the pattern to detect potential fraud.
• Image Scanning: Users can upload screenshots of payment receipts or chats. The AI uses computer vision (via Gemini API) to spot edited or fake details.
• Urdu & English Support: It understands both English and Roman Urdu/Hindi contexts, making it useful for local users.

⚙️ How we built it

I built this project using Python and the Streamlit framework for the user interface.

• The Brain: The core intelligence is powered by Google Gemini 1.5 Flash. I used this model because it is fast and efficient at understanding both text and images simultaneously.
• The Code: Written in Python using VS Code.
• Deployment: The app allows users to interact with the AI in real-time.

🚧 Challenges I ran into

As a Python learner, this was a challenging but exciting journey.

1. API Integration: Understanding how to securely pass API keys and make requests to Google Gemini was a big learning curve.
2. Multimodal Input: Handling both text and image inputs within the same interface required careful logic.
3. Deployment: Learning how to upload my code to GitHub and manage requirements.txt for the first time was difficult, but I mastered it!

🏅 Accomplishments that I'm proud of

• Successfully building my first AI-powered application.
• Learning how to use Generative AI for solving real-world problems.
• Publishing my code on GitHub professionally.

🚀 What's next for AI Scam Detector

I plan to add more local languages (like Pashto and Sindhi) and improve the detection accuracy. I also aim to turn this web app into a mobile application so that everyone can have a scam detector in their pocket.

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for AI Scam Detector

Built With

• code
• google-gemini
• python
• streamlit
• vs