Privacy & safety
Mentra holds some of the most personal things people say. We designed it so patients stay in control, the care team sees only what it needs, and anything worrying reaches a human.
Patients decide what's shared
Entries are private by default unless a patient chooses to share them, one at a time or all at once. Briefs are written only from shared entries.
Encrypted per patient
Entries and recordings are encrypted at rest, each patient's under a separate key. A key is only unwrapped for someone allowed to read that patient's data.
Scoping in the database
Every request is limited by organisation, by care assignment and by the patient's consent, and the database itself enforces those rules.
Staff see status, not content
Organisation admins, front-desk staff and Mentra's own operators can't open entries, briefs or flags.
Every clinical read is logged
Each time a provider opens clinical content, it's recorded. Patients can see who viewed their entries and when.
Deletion means deletion
Patients can delete an entry or their whole account. An organisation's retention rules apply only to that organisation's data.
Safety
When something's wrong, a person finds out.
A fixed crisis card
Crisis resources, including 988, are always one tap away in the app. The wording is written by people and never generated.
Flags go to a person
When an entry contains concerning language, it's flagged to the care team to review. Rules and a model both check, and either one can raise a flag.
No AI advice, no diagnosis
Mentra never replies to patients, never offers crisis advice and never diagnoses. Briefs describe what was said, cited to dated entries.
Mentra is not an emergency service
If you or someone else is in immediate danger in the US, call 911. For the Suicide & Crisis Lifeline, call or text 988. Safety flags are reviewed by a person during normal care, not monitored around the clock.
Questions about privacy or security?
We're happy to walk your compliance or security team through how Mentra works in detail.