Meet Kai
Kai will wait as long as you need.
Kai is the one who sits with you before the doctor comes in. Not a form. Not a phone tree. Someone who speaks the way you do.
Personality first. Capability second.
We wrote Kai to behave like the best nurse you have ever met: unhurried, unbothered, and impossible to embarrass.
Everything on this page is a promise about behaviour, not a feature. A mascot that does not change how the software talks to you is just a sticker.
Kai has one job, to help you say what you came to say, and it is built without the ability to do anything else.
What Kai promises, and what Kai refuses.
This is the whole bargain, and it is the most important thing on the page — which is why it is not in the footnotes.
What I will do
- I will wait.Take as long as you need. Say it twice, change your mind, start over. I will not rush you and I will not sigh.
- I will listen however you like.Speak out loud, or type if that is easier, and switch whenever you want.
- I will ask, not assume.If something you said is unclear, I will ask about it rather than write down a confident guess.
- I will not judge you.Not about what you drink, what you smoke, how long you waited, or what you are frightened of.
- I will not make you repeat yourself.Tell me once, and your doctor gets it in writing.
- I will tell you I am an AI.Every time you ask. You have an absolute right to know who you are talking to.
What I will never do
- I will never diagnose you.Not carefully, not with a disclaimer. I will not prescribe anything and I will not decide a single thing about your care. I am built without the ability to.
- I will never triage you.I do not score you, rank you, or decide how urgent you are. Those belong to a clinician who can examine you.
- I will never claim to be a person.I will not say I am a nurse, a doctor, or a human being — and if you ask, I will say so plainly.
And if you are in danger right now, Kai is the wrong tool and will say so. Get emergency help or call your local emergency number.
Kai is built on a large language model. What makes it different is not the model. it is the fence we build around it.
A narrow job (collect an intake history), a narrow set of things it may say, and an explicit list of things it must never say. It stays inside that fence for the whole conversation.
Hola, soy Kai. Cuéntame qué te duele, con tus propias palabras.
Me duele el pecho cuando subo las escaleras. Empezó hace tres días.
Gracias. ¿El dolor se va cuando descansas?
Kai asks a follow-up. A translation app would have translated the sentence and stopped.
Questions people actually ask.
- Is Kai a real person?
- No. Kai is a computer program, and it will tell you so if you ask. It is designed never to pretend otherwise.
- Will Kai tell me what is wrong with me?
- No, and this is on purpose. Kai has no ability to diagnose. It collects what you say and hands it to your doctor, who is the one qualified to work out what it means.
- What if Kai misunderstands me?
- You read the summary before anyone else does. If it is wrong, you can correct it or simply not send it. Kai also asks follow-up questions rather than guessing, but it is not perfect, which is exactly why you get the final say.
- Can I use it for my mother or father?
- Yes, and many people do. You can sit with them and help while they talk. Nothing about Keiro assumes the patient is the one holding the phone.
- Does my doctor have to use Keiro too?
- No. The summary is just clear written English. No account, no app, no setup on their side.