SafePay AI

"From QR Scan to Legal Insight — All in One Streamlit-Powered Pipeline."

✅ Inspiration

While working on a web scraping project in a previous hackathon, we realized how unstructured and overwhelming financial data can be—especially for small merchants and users trying to make informed payment decisions across borders. In the rapidly evolving fintech space, international transactions carry hidden legal, compliance, and fraud risks that often go unnoticed until it's too late.

That’s when we created SafePayAI — an AI-powered tool that empowers users and merchants to navigate the global payments landscape safely, by combining Generative AI-driven insights with real-time fraud detection.

✅ What it does

SafePayAI is a cloud-deployed fintech dashboard that helps users and merchants engage in secure, borderless financial transactions. It uses Google's Gemini AI to summarize global financial and legal news, flagging potential risks that could impact international payments or partnerships.

It also features a QR code fraud detection module that scans UPI and similar payment QR codes — helping users avoid phishing scams or fraudulent requests, including those from international sources.

Whether you're a small merchant receiving global payments or a user exploring cross-border financial opportunities, SafePayAI helps you stay informed, protected, and globally connected.

🛠️ How we built it

Frontend: Built using Streamlit for an intuitive and responsive user experience.

AI Integration: Leveraged Google Gemini API for summarizing legal and financial news and assigning a risk score.

Fraud Detection: Implemented QR code decoding using Pyzbar and PIL to scan UPI or foreign-looking payment codes.

Cloud Deployment: Deployed on Google Cloud Run using a containerized architecture to ensure scalability and low latency.

Security: Managed sensitive keys using .env and followed best practices in API and container security.

🧩 Challenges we ran into

Handling rate limits and prompt tuning with Gemini API to get consistent, actionable summaries

Dealing with environment variables and dependency issues during Google Cloud Run deployment

Optimizing image processing and responsiveness for real-time fraud checks

Making sure the tool was relevant globally, not just in an India-specific payment context

🏅 Accomplishments we’re proud of

Successfully integrated Generative AI for real-time global risk analysis

Built a working QR code fraud detection system for payment safety

Deployed a fully-functional, scalable AI app on Google Cloud

Aligned fintech security with global merchant needs, helping build trust in cross-border payments

📚 What we learned

How to effectively integrate LLMs into real-world fintech tools

Key deployment and containerization practices for Google Cloud Run

The role of risk visibility and compliance in enabling safe, borderless commerce

The power of combining multiple features — legal, financial, AI, and QR tech — into a single useful platform

🌍 Why it fits the theme

In a world moving toward borderless digital commerce, SafePayAI empowers users and merchants with the knowledge and protection they need to confidently transact across countries. From risk-aware financial insights to fraud-resistant QR payments, it reduces the friction and fear associated with global money movement.

SafePayAI isn't just about safer transactions — it's about building trust in the future of cross-border fintech

🌍 Future Prospects SafePayAI has strong potential to evolve into a full-stack, intelligent payment assistant. In future iterations, we envision:

✅ Real-Time Verified Payments: After fraud detection, the system can initiate secure, real-time payments to verified recipients—streamlining the transaction process.
✅ Multi-Agent AI Orchestration: Adding another layer of AI agents to handle tasks like compliance checks, currency conversion insights, and transaction optimization.
✅ Global Payment Intelligence: Expanding the platform to support region-specific payment systems and regulations, making it a truly global fintech companion.
✅ User Personalization: Tailoring risk alerts and financial summaries based on user profiles, transaction history, and geolocation