Skip to main content

zkAGI in plain words

This page explains what we do without technical language. It is written for patients, bank customers, citizens, journalists, and anyone else whose data ends up inside systems like ours. You do not need to know anything about AI to read it.

The problem, in one story​

Imagine your hospital wants to use AI to read scans faster. Helpful, right? But to do that, most AI services need your scan sent to a computer owned by someone else, in a place you will never see, under rules you never agreed to. The hospital gets a promise that your data is safe. A promise is all it gets.

Promises are not proof. If something goes wrong, nobody can show what actually happened to your scan. That is the gap we work on.

What zkAGI does​

We build the plumbing that lets hospitals, banks and law firms use AI without sending your data away.

Three simple ideas make it work:

1. The data stays home. The AI runs inside the hospital's own walls, or inside a boundary the hospital draws on a map. Your scan does not travel to a stranger's computer.

2. The hospital picks what to trust. Some organizations trust special sealed computer chips, the way you might trust a tamper-proof lock. Others trust mathematics, the way you might trust that a shredded document cannot be read. Some want both at once. We do not make that choice for them. Their own experts and lawyers pick, and the system enforces the choice.

3. Every use leaves a receipt. Each time the AI runs, it produces a receipt. Like a shop receipt, it records what happened: what ran, where, and under which protections. An inspector can check that receipt later without taking anyone's word for it. That is the difference between "trust us" and "check for yourself."

What this means for you​

  • If your hospital or bank uses infrastructure like this, your records can get the benefits of AI without leaving their custody.
  • If a regulator or auditor asks "what happened to this person's data?", there is a receipt to show, not just a policy document.
  • Nobody, including us, has to be blindly trusted. The protections are chosen by the people responsible for your data and can be checked afterwards.

What we do not do​

  • We do not sell or share the data that flows through the system. It is not ours and we cannot read it.
  • We do not work on weapons or surveillance. The platform is built for peaceful, civilian use: medicine, banking, law, research.
  • We do not claim everything is finished. Parts of this platform are still research. Every page in these docs carries an honest label saying whether that piece is live today or still being built.

Want to go deeper?​

  • A gentle next step: Trust domains explains the "pick what you trust" idea properly.
  • The receipts page shows what our receipts contain.
  • Questions this page did not answer? Talk to a human. A real person reads every message.