How HMARA works
HMARA is built around a small number of things it actually does, not a long list of features it might one day have. This page describes exactly what happens when you use it, what it deliberately does not do, and how it handles what you share.
What actually happens when you use HMARA
Voice and text conversation
Speak naturally in voice mode or write in text chat — both go into the same conversation, and you can switch between them at any time.
Memory of relevant context
HMARA can recall relevant details and themes from earlier conversations, so you do not have to re-explain the same situation every time you open the app.
A tone you choose
Pick how HMARA responds — supportive, purely listening, encouraging, reflective, or structured — and change it at any time from your profile.
Personal insights
Recurring themes and connections surface from your own conversations, not from a generic questionnaire.
A wellbeing trend over time
A simple trend shows how things have been moving over recent days and weeks, so change is easier to notice than it is in the moment.
How your data is handled while this happens
Two rules govern everything above: your conversations are used to support you, not to train a model that anyone else sees, and voice input is converted to text and then discarded in its original audio form.
- Your dialogues are never used to train HMARA’s or any third party’s general models.
- Source audio is discarded once it has been transcribed to text; the transcript is what HMARA keeps.
What HMARA deliberately does not do
HMARA is built to be honest about its edges rather than to look more capable than it is. It does not:
- give a medical or psychological diagnosis
- provide therapy, counselling, or any other licensed clinical service
- replace a doctor, psychologist, psychiatrist, or other qualified professional
- act as an emergency service or a substitute for emergency responders
- produce a clinical score or a formal mental-health rating