Inspiration
Being university students, we have had our fair share of scam emails and messages. Oftentimes, there will be messages that pose an unclear risk. We decided to build a tool that will help bridge that uncertainty.
What it does
The user can type suspicious text into the text box, and the Nemotron AI will scan, analyze, and detect whether the text is at high, low, or suspicious risk. We included a fallback: when Nemotron fails, Claude AI will analyze the text instead. ElevenLabs text-to-voice is also integrated to assess messages.
How we built it
We first outlined the structure of our project. We then designed the web UI on paper, and since we both have limited experience with web dev, we decided to vibecode the UI with Claude to build its functionality.
Challenges we ran into
We had issues connecting the API keys. For one, Nemotron often wouldn't work, so that is why we included a Claude API fallback.
Accomplishments that we're proud of
We were proud of learning more about the web development process and being able to connect to different AI models using the API keys.
What we learned
We learned a lot more about JavaScript and connecting our project with different AI models.
What's next for Scan No Scam
If we had time to learn, we thought it would be best to make our application and extension that can easily scan emails or text messages automatically for you.
Built With
- claude-api
- elevenlabs
- javascript
- nemotron

Log in or sign up for Devpost to join the conversation.