Inspiration

Fact-checked something I kept seeing in a video: hospitals are legally required to post their real prices publicly, have been since 2021, applies to every hospital regardless of ownership. That part's actually true.

But there's no single "real price" for a procedure. Same thing can have three different numbers — the chargemaster rate (inflated, nobody pays it), the rate your insurer negotiates, and a separate cash-pay rate if you're uninsured. Which one ends up on your bill can mean paying 2-5x more than someone else for identical care. And uninsured patients, who have no insurer negotiating for them, usually get hit with the highest number.

BillcleaR exists because that data is sitting in massive government-mandated files that nobody can actually read.

What it does

BillcleaR does three things:

Hospital price lookup: search a hospital and procedure, see the real gross charge, cash price, and negotiated rate range, pulled directly from that hospital's own federally-mandated price transparency file. Bill translator: upload a photo or PDF of a real hospital bill; Gemini extracts every line item and explains it in plain language, cross-referenced against that hospital's own posted prices so you can see exactly where you land compared to what the hospital says it accepts. Procedure/medication explainer: look up any procedure or medication and get a plain-language explanation of what it is, why it's done, and side effects.

The goal isn't just information: knowing the cash price is genuine negotiating leverage. Studies show hospitals routinely accept 40-60% of billed charges from uninsured patients who negotiate directly, and cash prices are sometimes even lower than what insurance companies negotiate. Most people don't know any of this is negotiable at all.

How I built it

Backend in Python/FastAPI. Hospital price data — 11 Houston-area hospitals — is parsed directly from real CMS-mandated machine-readable files and normalized into Tiger Data (Postgres). Gemini powers bill extraction and plain-language explanations and orchestrates the app as an agent that decides which tool to call (extract a bill, look up a price, explain a term) rather than following a fixed script.

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for BillcleaR

Expanding hospital coverage beyond Houston, adding multilingual support (the same Gemini prompts are already structured for it), and a "what to say when you call billing" script generator built from the same parsed bill data; since knowing the numbers is only useful if you know how to use them.

Built With

Share this project:

Updates

Submission history