Rexpense

Inspiration

In this day and age where younger and younger generations are gaining access to individual spending, a gap opens where many have trouble regulating, or even understanding, where their money is going. This leads to continued poor money habits as they grow. By creating an application like Rexpense, bad habits can be caught early and corrected before it starts to influence their later finances.

What it does

Rexpense takes a picture of your receipt and utilizes AI to sort the expenditures into various categories such as grocery, shopping, restaurants, and misc. Targeted towards younger generations, Rexpense can help people recognize how their spending money both in and out of their bank accounts, making it perfect for younger audiences who may not have access to a traditional bank.

How we built it

Frontend was built using React Native, while Backend was build using Node.js. We chose this tech stack for facilitated integration. Backend used Tesseract to parse data from the image uploaded from frontend, after which we used the SambaNova Cloud API (based on the Meta-Llama-3.1-405B-Instruct AI model) to return specialized information back to the frontend. This whole flow was successfully accomplished using a RESTful API structure.

Challenges we ran into

Merging branches when integrating frontend and backend. Additionally, figuring out how data from images would be directly communicated with the SambaNova Cloud API was challenging.

Accomplishments that we're proud of

We have a working camera built into the app that allows for easy receipt uploads. The AI API is integrated successfully, leading to accurate interpretation of images/receipts.