Zentra

Find the credit cards that work for you.

Comment

✨ Inspiration

Credit cards come with endless options, confusing terms, and hidden trade-offs. Choosing the right one often feels more overwhelming than rewarding. We wanted to create a tool that makes this process simple, helping people find the cards that match their needs and giving them a clear path to maximize the benefits available to them.

💳 What It Does

Zentra is a personalized credit card recommendation platform that makes choosing the right card simple and stress-free. Users begin by completing a short survey about their spending habits, lifestyle, and financial goals. Their responses are processed by our backend and analyzed through the Google Gemini API, which compares them against a structured database of credit cards.

Instead of overwhelming users with endless options, the system narrows the results to the top three credit cards that best fit their profile. Each recommendation includes key details such as rewards categories, fees, and perks. From there, users can save their preferred cards to their personal account for quick reference later. The platform combines secure authentication, streamlined design, and tailored recommendations so that, in just a few clicks, users can discover and keep track of the cards most likely to benefit them.

🛠️ How We Built It

Our backend is powered by FastAPI in Python, which manages survey processing, authentication, and recommendation logic. We integrated Google Gemini to analyze user responses and generate tailored credit card matches. MongoDB Atlas stores card data and user profiles, while Supabase handles authentication and secure account management.

On the frontend, we built a responsive interface using React, TypeScript, and Tailwind CSS, ensuring a clean and intuitive user experience. The frontend communicates seamlessly with the backend through custom API endpoints, allowing users to view their top three recommended cards and save their selections for easy access later.

❗ Challenges We Ran Into

A major hurdle was dealing with fragmented and inconsistent credit card data. We had to scrape, clean, and restructure information from multiple sources before it could be used effectively in our backend and recommendation engine. Another challenge was designing the flow of information between the frontend, backend, and Gemini API. It took several iterations to decide how to send survey responses and format recommendations so that the system stayed smooth and reliable. Along the way, we also experimented with different authentication providers. We began with Auth0, but recurring integration issues led us to pivot to Supabase, which allowed us to keep authentication secure while moving faster during the hackathon.

🏆 Accomplishments

We are proud to have created a simplified financial pathway that adapts to each user’s specific needs and preferences, helping maximize credit card rewards and long-term benefits. A key accomplishment was learning and applying technologies we had never worked with before, including FastAPI, TypeScript, and the Google Gemini API, under tight time constraints. We successfully integrated multiple systems, from survey input to backend processing and AI-powered recommendations, delivering a seamless experience in just 36 hours. Thanks to our flexible infrastructure, most of our features are horizontally scalable, allowing the system to handle increased users and data with minimal changes.

📚 What We Learned

During the hackathon, we gained hands-on experience with FastAPI, TypeScript, and the Google Gemini API. Beyond technical skills, we learned how to integrate APIs effectively, structure frontend-backend communication, and design a recommendation system that is both user-friendly and scalable.

🚀 What's Next for Zentra

We plan to expand Zentra with more data sources, smarter AI recommendations, and enhanced tracking of saved cards. Future improvements include alerts for new offers, a smoother interface, and mobile support, all built on our flexible and scalable infrastructure ready for growth.

Built With

Share this project:

Updates