Credit Daddy

You've got questions about your potential credit card. We can answer them with AI.

Comment

Inspiration

Reading credit card cardholder agreements is a highly taxing process and involves the possibility of missing key details. It was also difficult to ask questions about the language in the convoluted contracts, and alternative credit cards often existed. Financial literacy shouldn't be a barrier to financial well-being, and we wanted to make learning about credit cards easier with the help of Cohere AI with Credit Daddy. If GoDaddy could make purchasing a domain accessible to others, Credit Daddy would make learning about credit cards accessible to all.

What it does

Users upload their cardholder agreements to the Credit Daddy website in PDFs. Credit Daddy uses AI to read the document and give glanceable information about the card, including average APR, annual fees, overcharge fees, and more. While viewing information about the card, users can ask questions about the card with AI-generated responses and a dynamically generated FAQ section based on other user's questions.

Finally, users are presented with the option of comparing the credit card they currently view to other credit cards. All credit card data is based on AI-generated data points from credit cardholder agreements.

How we built it

Credit Daddy is a web app where we used Python as our backend along with Cohere AI to comprehend the cardholder agreement PDFs and answer users' questions. The web app uses a Flask API for data processing and HTML, CSS, and Javascript to design and build our web app. Our database of choice is Google Firebase Firestore for efficient retrieval of data.

Challenges we ran into

Some challenges included getting all members up to speed with our development tools and environment. We also needed help to get our Figma design to work on our HTML and CSS web app. In addition, it was challenging to learn how to collaborate on GitHub with environment variables, merge conflicts, and get started on Python. Getting our API to work proved to be difficult, but mentors were able to assist us in identifying bugs and recommending improvements. Finally, Cohere AI proved challenging to work with, including context limits below our PDF sizes and understanding which model to use.

Accomplishments that we're proud of

We're proud of successfully shipping our project by the deadline onto the internet! Although everyone had different strengths, we were able to find ways to contribute and launch an ambitious product while delivering a beautiful user experience. Completing the project involved many late nights, seeking mentorship from others, and stepping outside our comfort zone with new uses for Python and Git best practices.

What we learned

We learned how to develop a full-stack web application using Python as a back-end and the Cohere AI API. In addition, we learned to design beautiful user experiences while maintaining a feature-rich product.

What's next for Credit Daddy

We plan to expand Credit Daddy to more users, add more credit cards to our database, and expand the feature set to cover more use cases, such as web-scrapping credit card websites. In addition, we hope to add user accounts to the web app and provide a more personal experience.

Built With

Share this project:

Updates