BillBuddy

IBM x RUBITS Datathon Project

BillBuddy is an AI-powered healthcare bill negotiation copilot that helps patients understand and challenge their medical bills. Users upload a photo of their bill, and BillBuddy automatically reads it, compares the charges against real hospital pricing data, flags any overcharges, and generates a personalized negotiation script so they can confidently fight for a fair price.

TEAM

Name	Role
Ali	Backend Lead (Flask API)
Rushil	Data & ML Engineer (PyTorch)
Antonio	Frontend Developer (UI)
Taha	Infrastructure / DevOps
Thomas	Product / UX / Documentation

Procedure

1, Upload an image of a medical bill.

2, OCR will convert the image into machine readable text.

3, BillBuddy will parse hospital, procedure, amount and balance.

4, BillBuddy will then compare the prices against NJ hospital price dataset.

5, BillBuddy will predict expected price with PyTorch model.

SETUP

1. Clone the repo

   git clone https://github.com/<your-username>/BillBuddy.git
   cd BillBuddy

2. Create a virtual environment

python -m venv venv

Activate it:

• Windows: venv\Scripts\activate
• Mac/Linux: source venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Run the backend

cd backend
python app.py

You should see Flask start up at: http://127.0.0.1:5000/

Visiting that URL in your browser should return:

{ "status": "BillBuddy backend running" }

5. Open the frontend

Open http://127.0.0.1:5000/ in your browser. The backend serves the frontend; do not open frontend/index.html directly (that can cause "Failed to fetch").