MEET
Leo2.0
For the questions that come up after you leave the appointment.
PRIVATE AI HEALTH SUPPORT
RUN INDEPENDENTLY BY PROJECTLEO
Lab results and chart notes don't always make sense the first time you read them. By the time you think of the real question, the appointment's over. Leo 2.0 gives you somewhere to go back to it, privately.
Sound familiar?
“These test results might as well be in another language.”
“I had a dozen questions in the waiting room. I forgot half of them once I was in the room.”
“This denial letter just says no.
No explanation. No next steps.”
Who's on the other end?
Before you share personal info into an AI tool, you deserve to know who sees it and where it goes.
ProjectLeo takes $0 from pharmaceutical companies, hospitals, or insurers.
We run independently of other AI companies.
We built Leo and we operate it ourselves.
Most health apps are paid for by someone else, a hospital system, an insurer, or an employer plan covering the bill, and that can shape what gets recommended to you. We looked for another health AI company with none of those ties that also doesn't sell patient data. We didn't find one.
Your conversations with Leo, and the health information you share, are stored only on ProjectLeo's own servers. We don't sell that information. Not to advertisers, not to insurers, not to anyone.
Everything you share is encrypted in transit and at rest, the same standard hospitals and banks hold themselves to. We also run regular security audits to keep it that way.
What went into building Leo2.0?
Leo2.0 has one job.
To help you make sense of what happened at the doctor's office after you're already home.
Practice conversations covering real situations
We generate practice conversations for Leo to learn from, covering people without insurance, caregivers, people with disabilities, and people dealing with cost or language barriers. We control which patients show up, so it doesn't default to the easiest cases. None of it comes from real patient records or real chat logs.
A separate check on every training example
A different model grades every practice conversation. We don't let the same system grade its own work. A person checks a sample of those grades by hand. We're looking for plain language, honesty about what Leo doesn't know, and answers that stay inside Leo's role. Anything that tells someone to change their treatment gets thrown out.
Testing for the hard questions
We test how Leo responds to hard and urgent situations, including moments when someone needs help right away. If a response fails, we fix it and test it again before it ships. This is our own process, not a government approval or a clinical certification. Leo can still get things wrong.
Why I built it to work this way
“I spent a lot of my childhood translating doctor's appointments and insurance letters for a parent who didn't speak English well.
I learned early what it costs a family when nobody explains anything in a language they understand.
That's why I built Leo2.0: somewhere private to ask the question you didn't get to ask in the room. I wanted you to know exactly who's running it.
ProjectLeo operates on its own, your conversations stay private, and we don't sell your data.”
- Charlie Alfarah, Founder of ProjectLeo
Bring Leo2.0 your next question
A lab result you don't understand. A line in your chart that doesn't sound right. A question you thought of on the drive home. Ask Leo2.0.
Leo provides health information and guidance, but does not replace professional medical advice, diagnosis, or treatment. Always consult a qualified healthcare provider about your health concerns.