This disclaimer applies to every AI-powered feature in HAVYT, including:
It works together with the Health & Medical Disclaimer, which applies fully to all AI output.
AI language models generate responses from statistical patterns, not from verified understanding of your situation. That means AI output can be:
Please apply your own judgement to everything Emelie says, and verify anything important with a qualified source before acting on it.
Nothing Emelie or any AI feature produces is medical advice, diagnosis, clinical assessment or treatment guidance — no matter how specific or personalised it sounds.
Never start, stop or change medication, prescribed therapy or any medical treatment based on anything in HAVYT. Those decisions belong with you and your doctor, rheumatologist, pharmacist or other qualified professional. Emelie is deliberately designed to point you towards professionals for medical questions; if she ever appears to do otherwise, that is an error — please report it (Section 10).
Emelie is not an emergency service and cannot reliably recognise or respond to a crisis — medical or mental-health. If you are in an emergency situation, or having thoughts of harming yourself, contact your local emergency number (112 anywhere in Spain and the EU), a crisis line, or a healthcare professional directly. See the emergency section of the Health & Medical Disclaimer for warning signs that need urgent care.
To make her answers relevant to you, Emelie is given context from your HAVYT data when you use AI features — for example your profile and goals, health conditions you have declared, recent check-ins and symptom logs, workouts, supplements, cycle information (if you track it) and connected-device signals. That is what "personalised" means here: Emelie reads your logs, not your mind.
Two honest consequences of this:
Details of what data is used, and your controls (including turning AI features off), are in the Privacy Policy.
AI processing for HAVYT is performed by OpenAI (OpenAI, L.L.C. / OpenAI Ireland Ltd for EEA users) acting as a data processor for HAVYT. When you use an AI feature, the relevant context described in Section 5 — and, for voice features, your audio; for photo features, your photos — is transmitted to OpenAI to generate the response.
Under OpenAI's API terms, API data is not used to train OpenAI's models by default. Full processor details, storage locations and retention are in the Privacy Policy.
HAVYT's pattern features (and Emelie, when she talks about them) describe statistical correlations found in your own logs — for example, "on weeks with less sleep, you tended to log more stiffness."
A correlation is an observation, not an explanation. It does not mean one thing caused the other, and it is not a scientific finding or a medical conclusion about your condition. Patterns are meant to be interesting starting points for your own reflection — and, if they seem meaningful, for a conversation with your healthcare professional.
If you choose to use photo analysis:
Photo analysis only runs when you actively request it, and you can delete your photos at any time (see the Privacy Policy).
When you talk to Emelie by voice, your speech is transcribed and the transcript is saved into your chat history, just like a typed conversation, so you and Emelie can refer back to it. The audio itself is processed in real time to enable the conversation and is not kept by HAVYT. Storage, retention and deletion of chat history are described in the Privacy Policy.
If Emelie says something wrong, unsafe, out of character, or that sounds like medical advice, please tell us. Reports like this directly improve the app's safeguards.
Email: support@havyt.app
Email: support@havyt.app
Provider identity, governing law and jurisdiction are stated in the Legal Notice (Aviso Legal) and the Terms of Service.