Every AI company says "trust us." We built you a button instead.
One click runs a live audit of Hush AI from inside your own browser, against the production system, right now. Not a report we wrote about ourselves: checks your browser performs and you can reproduce from any terminal. If a check ever fails, you will see it before we do.
What this can and cannot prove. Your browser can verify everything on the wire: the live zero-knowledge contract, the security headers, that this page loads nothing from anyone but us. It cannot see inside our servers; no company's audit button can, whatever it claims. So the wire contract is published at /v1/privacy-manifest and this page holds us to it in public, every claim is restated in our terms where it is legally binding, and pilot customers get a signed audit export of their own usage to check against their own records. Catch us lying and you will have caught it in writing.
What the button runs
Seven checks, all same-origin, all readable in this page's source (View Source, no build step, no minifier). Each result card carries the exact curl command that reproduces it without trusting our JavaScript.