ScamNoMore mobile app

• 

  ScamNoMore pic

Inspiration

Scams are on the rise in Singapore. The ScamNoMore app aims to help people identify any red-flag indicators easily by leveraging AI/LLM's capability in scam-pattern recognition. The SPF ScamShield lacks certain important features, such as local language feature (Chinese, Malay, Tamil), a powerful chatbot and sharing info among community/communities which ScamNoMore app has therefore incorporated all these features.

What it does

ScamNoMore is a mobile-first scam-prevention app for Singapore. Users can:

• Check a suspicious message by uploading an image or taking a picture.
• Receive AI-powered scam analysis, warning signals, and safety advice.
• Ask Eliza, an interactive scam-prevention assistant, for guidance.
• Search verified scam cases by scam type, town, keywords, and time period.
• View scam activity on a Singapore town heatmap.
• Report incidents to help identify recurring scam patterns.
• Join community rooms to share privacy-safe local information.
• Switch between English, Chinese, Malay, and Tamil.

How I built it

The app was built using Codex which created a responsive HTML, CSS, and vanilla JavaScript frontend, supported by a Python server and SQLite database. The database is seeded from 5000 mockup scam case data based on real Singapore scam trend/mix. OpenAI's Responses API provides image analysis and Eliza's interactive Q&A experience. Browser image upload and camera capture allow users to check suspicious content directly from the Home page.

Challenges I ran into

I needed to ensure several features fit clearly on a small mobile screen without requiring unnecessary scrolling. Other challenges included positioning town labels accurately over the Singapore map, keeping Eliza's responses concise and product-specific, handling vision-analysis fallback states, and ensuring search results are displayed five cases at a time.

Accomplishments that I am proud of

I created a complete end-to-end experience covering checking, searching, reporting, and community support. The app combines AI vision analysis, a database-backed local scam search, multilingual support, an interactive chatbot, town-based community rooms, and a responsive mobile interface.

What I learned

I learned that an AI assistant needs accurate product context and live application data to give useful answers. I also learned the importance of clear navigation instructions, transparent fallback behaviour, privacy-safe community guidance, and designing every feature around the user's next action.

What's next for the ScamNoMore mobile app

Next, I would deploy the app publicly and actively promote the app to my fellow Singaporeans, improve moderation for community rooms, add user accounts and saved reports, expand official scam data sources, improve the Singapore map visualisation, and continue improving multilingual AI responses.

Built With

• browsercameraapi
• css
• html
• openai-api
• python
• responses-api
• responsive-design
• rest-api
• sqlite
• vanilla-javascript