For journalists, podcasters, and reviewers.

“Audra audits your medical bill against federal billing rules in 60 seconds and drafts a citation-heavy appeal letter you can edit and send.”

“About 4 of 5 American medical bills contain errors. Most people don't catch them because doing so requires learning CPT codes, NCCI bundling rules, and No Surprises Act fine print. Audra does it for you in 60 seconds, with citations.”

Available on request. Email [email protected].

Stats below reflect the Audra install base since public launch. Refreshed continuously \u2014 our /built-in-public page has the live numbers.

• $210B of overcharges on US medical bills every year (Kaiser Family Foundation).
• 1 in 5 hospital bills contains at least one billing error (JAMA Internal Medicine).
• 60 seconds \u2014 median time for Audra to OCR a bill, cross-reference every line against federal billing rules, and produce a draft appeal letter.
• No data sold or shared. Files are encrypted client-side (AES-256-GCM) before they reach our servers.

“We built Audra for the patient who got a $3,000 ER bill, called the hospital, and was told “that's just what it costs.” The math behind that bill is public; the bill is just wrong.”

“Hospitals make billing errors at scale. AI can read them at scale. We're betting consumers want the tools insurers and large employers already have.”

“Every finding cites a specific federal rule or state statute. We don't say “this looks high” \u2014 we say “this exceeds the CMS allowable for code 80053 by 28x.””

Direct downloads below. Use the icon for app references, the OG card for article headers, and the screenshots for product callouts.

For SVG vector logos, founder headshot, or higher-fidelity screenshots: email [email protected].

Please don't modify the wordmark, recolor the logo, or place it on backgrounds with poor contrast. Otherwise: we trust your judgment.

• Audra catches billing-rule violations (NCCI bundling, MUE exceedance, modifier misuse, upcoding) that human review usually misses — about 30% of bills our users upload contain at least one citable error.
• Bills are encrypted on the user's device before upload (AES-256-GCM with per-record envelope keys). We never train models on private user bills.
• The product is built for individual consumers — not hospitals, not insurers, not employer benefits programs. We're aligned with the patient, full stop.
• Federal regulation cited in every finding includes No Surprises Act, CMS NCCI edits, MUE policy, modifier rules, ACA preventive-care guidance, and state-level consumer-protection law.

• Practice medicine, law, or financial advisory. We surface billing-rule violations; we don't opine on clinical decisions.
• Sell, share, or aggregate user data with marketing partners.
• Take cases under active litigation or claims > $25K — those go to patient-advocate firms or attorneys.