为临床信任而建。

Most security pages ask you to take a great deal on faith. We would rather explain how things actually work, in language a clinician — not a security engineer — can follow and check. The starting point is simple: the less your sensitive data has to travel, the fewer ways it can go wrong. So we designed the whole system around keeping that data close to where it is created and used, and around making sure that whatever does move is something you have explicitly agreed to move.

By default, analysis happens inside your facility. There is no quiet relay to a remote service, no background upload of images, no copy of a record sitting on someone else's machine because that was the convenient way to build it. If a feature genuinely benefits from sharing something — and some do — that sharing is opt-in, it is limited to de-identified or consented information, and it is visible to you rather than buried in a setting nobody reads. The principle is that you should never be surprised by where your data has been.

Trust also depends on being able to look back. Our outputs are deterministic and reproducible: the same input produces the same result, and every step is logged so it can be retraced later. That matters in a setting where someone may, months on, need to explain exactly why a particular result appeared and what informed a decision. A tool that cannot account for itself has no place in clinical work, however clever it is. We would rather be predictable and explainable than impressive and opaque.

Security is never finished. Threats change, software changes, and the careful posture that is right today needs revisiting tomorrow. We treat trust as an ongoing practice rather than a box that gets ticked once: we design for it from the first line, we build to recognised standards for clinical-software quality and data protection, and we hold our own work to review rather than assuming it is correct. When we are still working toward a particular assurance, we say so plainly instead of implying more than is true.

We also believe security improves when more people are looking. If you find a weakness, we want to hear about it — responsible disclosure is welcomed, not treated as an embarrassment to be managed. And the relationship runs both ways: you stay in control of your data, you decide what is ever shared, and you can ask us, at any time, exactly how something works. The honest answer to a security question is worth more than a reassuring one, and that is the kind of answer we aim to give.