Inspiration
Splitting bills with friends can get surprisingly complicated. Someone pays for the food, another person orders something different, and shared items, taxes and tips make calculations even more confusing. We wanted to make this process simpler and avoid the usual back-and-forth when figuring out who owes whom.
What it does
FairSplit AI helps groups divide expenses without the hassle of calculating everything manually.
Users can:
- Enter bills manually or describe expenses in everyday language.
- Upload a receipt and extract its items and prices.
- Assign individual items and shared expenses to different people.
- Automatically calculate each person's share, including tax and tip.
- View a clear settlement breakdown.
- Track payments made through PayPal or mark payments completed offline.
How we built it
We built FairSplit AI using Python and Flask for the backend, with HTML, CSS and JavaScript for the interface.
Google Gemini powers the natural-language bill parser and receipt scanning features. It helps interpret expense descriptions and extract information from uploaded receipts.
For payments, we integrated the PayPal JavaScript SDK and PayPal Orders API using the Sandbox environment. The backend handles order creation and payment capture.
We also used GitHub for version control and Render to host the application.
Challenges we ran into
One of our biggest challenges was handling different ways of splitting expenses. A person might pay for an entire item, share another item with two people and split drinks among the whole group. Making sure taxes, tips and rounding were distributed correctly took careful attention.
Integrating AI with the existing bill calculation system was another challenge. We needed to ensure that information extracted by AI was converted into accurate bill entries.
Setting up PayPal Sandbox and deploying the application also involved troubleshooting credentials, environment variables and server configuration.
Accomplishments that we're proud of
We're proud to have brought multiple ways of entering expenses into one application. Users can manually enter a bill, describe it in natural language or scan a receipt instead of typing everything themselves.
We're also proud of the individual expense breakdown and settlement tracking, which make it easier to understand exactly what each person owes.
What we learned
We gained practical experience working with Flask, APIs, GitHub, deployment and integrating external services.
We also learned that using AI in an application involves more than getting a response from a model. The output needs to be structured, checked and connected to the rest of the application.
What's next for FairSplit AI
We would like to improve the payment experience, complete more extensive PayPal testing and introduce shared bill history across devices.
We also plan to improve receipt recognition, support more currencies and make the application useful for larger groups and frequent expense sharing.
Built With
- ai
- api
- application
- bill
- checkout
- css
- expense
- flask
- gemini
- generative
- html
- javascript
- language
- management
- natural
- ocr
- paypal
- processing
- python
- receipt
- sandbox
- scanning
- splitting
- web