From the first heartbeat to finding yourself again.
They tell you the baby is the size of a mango. They do not help you remember the question you thought of at 2am, organise three months of blood pressure readings before a ten-minute appointment, or notice that nobody has asked how the mother is doing since the birth.
About
The mother is not a secondary character.
Calm is a feature.
Privacy is part of the product, not a legal footer.
Support must continue after birth.
High-risk does not mean high-anxiety design.
Users control what is shared.
AI organises and explains; clinicians diagnose and treat.
Every alert must have a clear next action.
Every clinical claim must have provenance.
Success is measured in supported lives, not addictive engagement.
AI agents as skills
AI that organises and explains. Clinicians diagnose and treat.
Every agent is narrowly scoped, permission-aware and auditable. None of them can diagnose you, change your treatment, or soften a safety warning.
Narrow by design
Each skill has a versioned contract listing exactly which tools and which records it may touch.
Safety is deterministic
Clinician-approved rules decide urgency before and after the model runs. The model cannot override them.
Grounded in approved content
Answers come from medically reviewed material in your language, with the source shown.
Fully audited
Every run records the skill, version, locale, tools used and cost. You can see it. So can your clinic.