About artificial intelligence
AI is a tool, not a therapist. Our mentor uses generative models to accompany you, not to diagnose or decide for you.
What we do
The mentor is built on a hardcoded pedagogical framework. It doesn't improvise: it follows a supervised process.
There are automatic validators that block responses violating our principles (empty imperatives, validation without context, quick congratulations).
When the AI gets it wrong, we fix the prompt and document it in the public changelog.
The final decision is always yours. The mentor returns questions, not verdicts.
The mentor estimates your emotional state (emotion, energy, trend) from what you write, only to adapt how it accompanies you. You can see it in your panel, correct it by talking, and it disappears if you delete your account.
What we will never do
We will never use your conversations to train AI models — neither our own nor third-party ones.
We will never generate clinical diagnoses or medical recommendations.
We will never pretend the mentor is human.