ClearBill

“An AI-powered platform that analyzes medical bills, flags potential errors, and helps patients dispute unfair charges.”

Comment

Inspiration

Medical bills are one of the most confusing and stressful parts of healthcare. Many patients receive statements filled with vague descriptions, duplicated charges, and unclear line items, yet have no practical way to verify whether they are being charged correctly. Most people end up paying simply because they do not understand what to question.

I built ClearBill to give patients clarity, transparency, and confidence. Our goal was not to replace providers or insurers, but to empower everyday users with the ability to understand their bills and take informed action when something looks wrong.

What it does

ClearBill is an AI-powered web application that helps users analyze and challenge medical bills.

Users upload a bill (PDF, JPG, or PNG), and ClearBill:

  • Extracts text using AI-powered OCR
  • Structures the charges into individual line items
  • Analyzes each item using explainable, domain-informed rules
  • Assigns a Fairness Score (0–100) summarizing potential concerns
  • Flags issues such as:
    • Duplicate charges
    • Vague or unclear descriptions
    • Temporal inconsistencies
    • Illogical combinations of services
  • Allows users to ask questions through an AI chat interface (e.g., “Why is my score low?”)
  • Generates a professional dispute letter that users can download and send to billing departments

ClearBill does not provide medical or legal advice. It is designed as a patient advocacy and education tool that helps users understand what to ask about.

How I built it

ClearBill was built using a modern, full-stack architecture:

Frontend

  • Next.js with TypeScript for a fast, scalable web interface
  • Tailwind CSS and shadcn/ui for a clean, intuitive user experience
  • Deployed on Vercel for reliability and performance

Backend & Data

  • Supabase (PostgreSQL) for structured storage of bills, parsed data, analysis results, and scores
  • Supabase Storage for securely storing uploaded documents

AI & Processing

  • Google Gemini (Vision + Text models) for:
  • OCR (extracting text from bills)
  • Parsing raw text into structured line items
  • Analyzing billing patterns for potential issues
  • Powering the AI explanation and chat interface

System Flow

  1. User uploads a medical bill
  2. File is securely stored in Supabase
  3. Gemini Vision performs OCR to extract text
  4. Text is parsed into structured line items
  5. An analysis engine applies explainable rules to flag issues
  6. A Fairness Score is computed and stored
  7. Users can ask questions and generate a dispute letter

This modular design ensures clarity, scalability, and transparency in how each result is produced.

Challenges I ran into

  • OCR reliability: Medical bills vary widely in format and quality, making text extraction and parsing non-trivial.
  • False positives: We needed to balance between flagging meaningful issues and avoiding over-reporting. This required careful design of explainable heuristics.
  • Explainability: It was important that users understand 'why' something was flagged, not just that it was flagged.

Accomplishments that i'm proud of

  • Built a fully functional end-to-end system: upload → OCR → analysis → scoring → chat → dispute letter
  • Created a Fairness Score that summarizes complex billing data in an intuitive, visual way
  • Designed a clean, user-friendly interface that non-technical users can easily navigate
  • Integrated explainable AI, ensuring transparency instead of a black-box result
  • Delivered both frontend and backend as a complete, working project

What I learned

  • Applying AI to real-world documents requires careful handling of data quality and edge cases.
  • Explainability is just as important as accuracy, especially in sensitive domains like healthcare.
  • Good UX dramatically increases the impact of technical systems clear visual cues and structured explanations matter.
  • Building responsibly in healthcare means setting proper expectations, protecting users, and avoiding over-automation of decisions.

What's next for ClearBill

  • Integrating insurer Explanation of Benefits (EOBs) for deeper comparisons
  • Adding regional and regulatory context to strengthen dispute recommendations
  • Exploring enterprise use for patient advocacy groups and healthcare transparency initiatives

ClearBill’s long-term vision is to make healthcare billing more transparent, fair, and understandable for everyone.

Built With

Share this project:

Updates