Inspiration
Inspired by my friend getting scammed while using Cash App
What it does
Swivel Shield is a Cash App–inspired payment system that protects users from push payment scams by evaluating real-time transaction risk. When it detects potential fraud, it disrupts emotional urgency with spoken voice alerts, mandatory 3-second cooldown timers, and an optional 24-hour recallable Solana escrow safety net. For routine transactions, it provides a zero-friction fast track that completes instantly without warnings or security delays.
How we built it
Moving our development workflow to Google AI Studio was the key breakthrough that brought Swivel Shield to life. By leveraging Google AI Studio, we seamlessly integrated Gemini's real-time AI risk evaluation into our Cash App–inspired payment pipeline to instantly analyze transaction memos and flag coercive scam language. To make our auditory circuit-breaker fully functional, we generated custom, high-clarity voice alerts using ElevenLabs and engineered the frontend to seamlessly trigger the warning.mp3 file during high-risk transfers. This shift to cloud AI tools solved early environment bottlenecks and allowed us to deliver a complete, highly polished behavioral fraud defense system.
Challenges we ran into
During development, switching our primary environment to Google AI Studio was the project's turning point. Moving away from local configuration bottlenecks in VS Code let us quickly integrate and fine-tune Gemini's real-time risk evaluation API, streamline state management, and deploy our dual-path UX prototype seamlessly. This shift solved our integration challenges and enabled us to deliver a fully functional, production-ready payment security platform.
Accomplishments that we're proud of
We built a real-time defense system that stops payment scams in their tracks by cutting through emotional urgency using voice alerts, a mandatory 3-second cooldown, and a 24-hour Solana escrow safety net. Switching to Google AI Studio was a huge win for us—it helped us push past early setup roadblocks and launch a smooth app experience that blocks scams without slowing everyday transfers.
What we learned
We learned how to leverage Google AI Studio and Gemini API for real-time natural language risk evaluation, as well as how to integrate ElevenLabs audio assets into front-end state management. Through designing Swivel Shield, we gained deep insights into behavioral cybersecurity, discovering that effective scam prevention requires disrupting user psychology rather than relying on standard pop-up warnings that people routinely ignore.
What's next for Swivel Shield
We plan to fine-tune our risk evaluation models using real-time transaction telemetry and dataset insights from Swivel to continuously adapt to emerging social engineering tactics. From there, we will deploy Swivel Shield to production and expand our Solana smart contract infrastructure to support full-scale, on-chain escrow dispute resolution.
Log in or sign up for Devpost to join the conversation.