Nice Try

Inspiration

Scams are getting harder to recognize, and knowing the warning signs isn’t the same as responding under pressure. We wanted to give people a safe place to practice before facing a real scam.

What it does

Nice Try simulates scam phone calls and text conversations across six scenarios, including fake bank alerts, IT support, and family emergencies. Users interact with an AI scammer, then receive personalized feedback on warning signs, their responses, and safer ways to handle the situation.

How we built it

We built the frontend with React, TypeScript, and Vite, with a Y2K-inspired design. A Python Flask backend coordinates the training sessions and feedback. ElevenLabs powers the voice conversations, while Photon connects the text simulations through iMessage.

Challenges we ran into

Connecting the voice and text services took troubleshooting, especially around authentication and session updates. We also worked through live transcript scrolling, call controls, and keeping the phone and text experiences separate while sharing the same scenarios.

Accomplishments that we're proud of

We’re proud of bringing realistic phone and text scam training into one application, giving users two ways to practice handling suspicious conversations. Connecting ElevenLabs and Photon helped us turn our idea into an interactive experience where users can respond in their own words and see how the conversation unfolds. We’re especially proud of the personalized feedback, which turns each simulation into a learning opportunity by explaining the warning signs, what the user handled well, and where they could improve. We also put a lot of care into the Y2K-inspired interface, giving Nice Try its own personality while making the training feel approachable. Seeing those pieces come together into a working project over one weekend was a big accomplishment for our team.

What we learned

We learned how to connect multiple services into a working application and handle live conversation updates. We also learned how much small interface details—like scrolling behavior and button placement—affect the experience.

What's next for Nice Try

We’d like to expand Nice Try with more scenarios that reflect the scams people encounter in everyday life. We also plan to refine our ElevenLabs agents to make phone and text conversations more natural, varied, and responsive to the user’s choices. Improving the feedback is another priority: we want users to understand which warning signs they recognized, what they missed, and how they could respond more safely next time. Testing with more users will help us identify what works well and where the experience needs improvement. Our goal is to make Nice Try a practical, approachable training tool that helps people build confidence and feel more prepared when a suspicious call or message arrives.