Answered plainly.

A premium medical AI that brings your medical history into one place. It reads your records — lab reports, visit summaries, imaging reports — and combines them with your markers, your symptoms and everything you take into a single living health profile. Then it answers questions against the whole of it, in plain language, showing the data behind each answer.

Antelic is in early access and it goes out personally rather than through a sign-up form. Write to hello@antelic.com and say roughly what you’d bring to it — a panel, a stack, a history nobody has seen all of at once — and you get a way in.

Nothing during early access. Pricing has not been announced, and there is no payment system in the product today. When there is a price, it will be for the software rather than for anything you take, and the safety layer will never be behind it.

It is read once to extract the values. You see everything it found and confirm it, and only what you confirm is saved. The file itself is never stored and the extraction is never cached. The full sequence, written so you can check it against what the product does, is on Trust & safety.

No. Nothing you save is used to train anything, by Antelic or by anyone else. The model Antelic calls runs under a commercial agreement that excludes training on the requests sent to it, and Antelic itself trains no models and collects nothing for that purpose.

No, and it is built not to. Antelic organizes, explains and prepares you for a conversation — clinicians diagnose and prescribe. It never tells you what dose to take, it never ranks one brand against another, and where something looks urgent it says so plainly and tells you to contact a clinician.

It reads documents, explains what a marker means, and describes what tends to go with what. The strongest word it may use about a correlation is suggestive, and that limit is enforced in the code. It is never asked what dose you should take, it cannot raise or lower a safety flag — that layer runs with no model involved — and where the evidence is genuinely thin, it says so rather than filling the gap. Your weekly read is not written by the AI at all: it is computed directly from your own logs and your own results.

In a managed Postgres database, encrypted in transit and at rest, with every row scoped to your account by a rule enforced inside the database rather than a check in the app. Deletion is immediate and self-serve: delete your account from inside the product and your records go with it, without writing to anyone. You can also take your record with you at any time, in either of two forms: a complete printable copy, or a structured JSON file holding your profile, your stack, your logs and your bloodwork.

That question is with counsel, and answering it either way before then would be a guess about a legal status. What can be said now: there is no such thing as a HIPAA certification — no government body issues one — which is why no badge appears anywhere on this site. The controls Antelic applies regardless of the answer are listed on Trust & safety.

Antelic reads an Apple Health export file today and merges it into the same timeline as everything else, with no AI involved in the parse. Continuous sync from Apple Health, Oura, Whoop and CGMs is in development. Nothing on this site claims data Antelic doesn’t hold.

Not yet. Early access runs in your browser, on a phone or a desktop, and the desktop layout is a real layout rather than a stretched phone screen.

security@antelic.com. Send what you found and how to reproduce it. You get a reply from a person, and nobody acting in good faith will be pursued for reporting to that address.