Penni Personal Finance

Login screen
Create a new account button screen
Transfer screen within an account
Landing Page
Account page for a specific account
Asking Penni to do a transfer
Asking Penni about a user's spending habits
Asking Penni a simple finance question

Inspiration

We wanted to try and build something banking related, and also, we wanted to get some experience with building and using AI agents. We wanted to make a fully-fledged web app to try and get all aspects of development by having everybody on our team work on something different and eventually bring all the code together.

What it does

Penni Personal Finance is a web-based app that is meant for tracking a user's personal finances. It can track multiple bank accounts across different banks and also allows the user to see a brief transaction history on the main page. We have an AI assistant named Penni that is the main focus point, as Penni allows the user to talk to the AI agent in order to have Penni make transfers, deposits, or withdraws on their behalf. Penni can also provide some spending habit analysis based on the user's transaction history.

How we built it

We built it using HTML/CSS/JavaScript and Chart.js for the frontend, and Python, FastAPI, Uvicorn, OpenAI API, and JSON for the backend, database, and AI agent.

Challenges we ran into

One challenge we ran into was figuring out how to program our own AI agent and how to limit its responses to only what we wanted within our app. Another challenge was figuring out how to divide up the work evenly while still being able to learn. Integration of the backend, frontend, and the AI agent was also a big challenge as we tried to combine all of our work in order to make it run smoothly.

Accomplishments that we're proud of

We are definitely proud of building our own personal finance AI agent and being able to integrate the frontend and backend as well and making the web app work as intended. We are also proud of making the UI look nice for the whole project. Lastly, we are very proud of our Red Bull tower :)

What we learned

We learned how to create an AI agent, and how to limit the prompts the user submits and the outputs that the agent gives back. We learned how to interact with a simple server in order to store financial and account information, as well as how to design the frontend. The frontend took a lot of designing and planning as that is how we decided how the backend and AI would look. We overall also learned a lot about how to organize our code and how to use git workflows as a team. Lastly, we learned how to review code by hand and how to use AI to help us debug and find improvements in our code.

What's next for Penni Personal Finance

We would add better security features like voice recognition, passphrases, and maybe password verification. We would also add some more features to the personal finance app in general like adding credit cards, loans, and other investments. We would also just make the UI more in depth and make the database hosted so everyone can access it.