This page is for you, the patient, if your dentist records your jaw motion with Toothprint. It is short on purpose, and it is written to be read by you and not by a lawyer.
In plain words. Your dentist recorded a video of your face and a scan of your teeth. That material belongs to your clinical history and your dentist is responsible for it. We store it for them. We only use it to train our models if you signed a separate authorisation saying so, which you can refuse without anything changing in your treatment, and which you can take back later.
Your relationship is with your dentist. If you want to know what was recorded, to correct it, or
to have it deleted, ask them first. If they cannot help you, write to privacy@toothprint.ai and
we will route it.
To measure how your jaw moves, the session records video of your face while markers are held or attached near your mouth. That video is of you. It is not "video of the markers", and we are not going to describe it that way.
Alongside it: the scan of your teeth, which is as unique to you as a fingerprint; your name and identification as your dentist entered them; and whatever they note about your case.
We do train our own models on patient material. Only on the material of patients who gave a separate written authorisation for exactly that, ticked by them and by nobody else.
If you did not sign it, your material never enters a training set. The app will not even record you without your authorisation being registered first.
Things we are not going to tell you, because they would not be true:
Saying no changes nothing in your care. Nobody will treat you differently, and your dentist gets exactly the same tool either way.
If your dentist uses the feature that turns your photo into a stylised portrait for your record card, that photograph is sent to Google's Gemini API and comes back styled. Only the photo goes: not your name, not your scans, not your session video.
It only happens if you ticked that box — it is a separate box on the form you sign, and it starts off — and your dentist then presses the button. If you did not tick it, the server refuses to send the photograph and deletes it. You can change your mind at any time.
What Google keeps on their side is governed by their terms and not by ours, which is exactly why we are telling you rather than leaving it in a list of providers.
You may ask to know what we hold, to correct it, to delete it, to take it with you, and to withdraw an authorisation at any time. In Colombia these are the consulta and the reclamo of Ley 1581 de 2012, answered within 10 and 15 business days. In the European Union and the United Kingdom they are your GDPR rights.
Ask your dentist, or write to privacy@toothprint.ai. We verify who you are before answering,
because answering the wrong person would be the failure this is meant to prevent.
One limit, and it is not ours: your dentist must keep your clinical history for the period their own law requires — in Colombia, under Resolución 1995 de 1999. Deleting your account with us does not erase your clinical record with them.
Your dentist may ask you to wear a cheek retractor, to bite, to slide your jaw forward and to each side, and to chew. It usually takes a few minutes.
You may stop at any moment, for any reason, without explaining it. If anything hurts or worries you, say so; the procedure is your dentist's responsibility, not ours.
To Google's cloud in the United States, and to a GPU provider in the United States that performs the heavy reconstruction. Both are listed at Subprocessors. We do not sell your material to anyone, ever, and there is no advertising anywhere in this product.
These terms supplement the General Terms and the Privacy Policy, and they prevail over both on anything that affects you as a patient. They do not give you an account or the right to use Toothprint on your own.
In force since September 11, 2026 · Version 1.1
Version 1.1, amended 12 September 2026. Version 1.0 was published on 11 September and contained statements that a review against the running system showed to be wrong. They are corrected here rather than quietly edited, because a published document that changes without saying so is worth less than one that admits it changed: a page that said no third party ran AI inference on patient material, while a patient-photo feature was sending faces to a generative model; a security page that claimed point-in-time recovery and 12-month backups that were never configured; a deletion promise that gave one timeline for three destinations that do not run at the same speed; and a three-year destruction promise with no machinery behind it.