Patient memory
The patient memory is the structured record nexmin keeps of each therapeutic process: what every validated session contributes — recurring themes, open threads, active hypotheses, key events — compiled so the AI re-reads it before each new analysis. It is editable and the therapist's version wins: if the AI proposes a reading you do not share, you correct it and the memory keeps yours.
It has two layers. An episodic one: session fragments indexed so you can search by meaning, not just exact words. And a semantic one: the compiled portrait of the process — traits, open threads, hypotheses in progress, a timeline of key events. Pensa reads the second as context before looking at the new session, so the analysis arrives knowing what is already known. It exists for a practical reason: looking at one session is affordable; re-reading a thirty-session process in full, every time, is not. The memory compresses the trajectory so that reading the process between sessions is possible without reprocessing the whole history — and so the therapist, picking someone up after months, does not start from zero. Its limits are part of the design: it is only written from sessions the therapist validates, sensitive changes require explicit acknowledgement, and like all clinical content in nexmin it lives encrypted in the European Union and never trains models.
Inside nexmin
Each patient's record in the App: Pensa reads it as context on every analysis, and the therapist can edit it at any time — their version prevails.
lectura larga: la página completa sobre este tema →
Related terms
Last updated: 2026-07-28