The person behind the library.
Thylib is a small, independent product with a simple job: bring your scattered medical records into one place your own AI can read. No ads, no data brokering, no growth team — a subscription pays for the service, and that's the whole business model.
Why this exists
Thylib exists because of my kid — I'll call them Avery.
Avery has lived with depression for five years. They've improved, truly — but they still can't routinely go to school, even though they badly want to. If you've walked beside someone through something like this, you know what the appointments are like: fifteen minutes, always spent on the most urgent thing, never on the whole picture.
Earlier this year, in a long, unhurried conversation with an AI that holds the medical literature and has all the time in the world, I came across atypical depression— and it described Avery far better than anything we'd been working with for years. That distinction isn't academic: atypical features can change which treatments are worth trying, so it went straight onto the list for Avery's doctor. Then I went further and brought Avery's medical records into the conversation. They showed something I had quietly suspected but no single visit ever flagged: nearly every heart-rate reading in Avery's chart, across years of appointments, was high. It was a pattern that only becomes visible when someone — or something — can read the whole chart at once.
None of that is a diagnosis, and I treat none of it as one. What it did was make me a far more intelligent participant in Avery's care — for the first time, I walk into appointments with specific, informed questions instead of a knot of worry, and we spend the fifteen minutes on what matters.
Avery is not out of the woods. But I'm more hopeful than I've been in years, and I built Thylib because every family navigating something hard deserves the same chance to understand what's in front of them.
(Thylib starts with medical records, because that's where the most signal was for us. More kinds of health data are on the roadmap.)
Who's behind it
I've spent my career in healthcare — including about ten years at Kaiser Permanente, much of it working on chronic-disease management. That work taught me two things Thylib is built on: how much signal sits unused in medical records, and how rarely it reaches the fifteen-minute visit.
I keep my name off this page for one reason: the story above is my family's, and their privacy comes first — the same principle Thylib is built on. If you want to talk to me directly, the support address below reaches me.
Thylib is made and operated by Perceptive One LLC, a California company. It's the legal entity behind the Terms of Service and Privacy Policy — the party legally bound by every promise on this site, under California and U.S. consumer-protection and health-data law.
The promises, in one place
Your records are encrypted in transit and at rest, and are never sold, shared with advertisers or insurers, or used to train models on our side. The only place they go is the AI you connect, when you ask a question. You can delete everything, any time, and records are permanently destroyed within 14 days. Where the honest answer is nuanced — what your AI provider keeps, where coverage has gaps — we say so on the page, not in the fine print: Limits & accuracy.
Reach a human
Every email to customer-support@thylib.com is read by a person — usually me. Questions, a hospital that won't connect, an answer that looked wrong, or a deletion request (including from someone whose records a family member connected): same address.
Curious what it actually does before trusting it with anything? The live demo runs on a fully synthetic family — no account, no card, nothing real.