We're a small team trying to make medical billing suck less.

You upload a medical bill (PDF, photo, EOB, anything). We run it through a pipeline that combines OCR, federal billing-rule libraries (CMS NCCI edits, MUEs, modifier definitions), and an LLM trained on the patterns that cause real overcharges. Sixty seconds later, you have an audit with every flagged charge, a citation for the rule behind it, and a draft appeal letter ready to send.

If we don't find anything, the audit was free. That's the money-back guarantee on every one-time audit — the subscription tiers are predictable monthly costs designed for households with recurring care.

• Sell your data. Your bills + audit results are encrypted at rest with per-record keys. We will not sell, share, or aggregate them with marketing partners.
• Train models on your bills without explicit consent. Every model improvement we make uses synthetic data + the public, opt-in research dataset — never your private bills.
• Use dark patterns. Cancellation is one click. Refunds for unused audits are honored. No “are you sure?” loops with a hidden X.
• Practice medicine or law. Audra surfaces billing-rule violations; we don't give legal or medical advice. For complex disputes we'll point you at a patient-advocate firm instead of pretending we can handle it.

The team is remote across the US. The product is built and operated by a small group of engineers, designers, and a former hospital revenue-cycle analyst whose entire job used to be telling people their bills were correct. Spoiler: a lot of them weren't.

We're not VC-backed. We're funded by the subscriptions + one-time audits people pay us for. That constraint shapes everything else — we'd rather charge a real price for a real product than chase attention.

The fastest path is in-app support (/settings/support) — every ticket goes to a human, usually answered the same day. Otherwise: [email protected].