Resolut uses artificial intelligence in two places. This section states exactly what the AI is, what it sees, and what it does not.
Short supportive coach messages in the app are generated by Google's Gemini model (currently Gemini 2.5 Flash), called from our server. A separate internal evaluation adjusts which content the app surfaces to you; that evaluation runs on our own rules over your in-app signals, and its result is never shown to you as a label.
Every coach request is assembled server-side from a fixed, bounded set of fields:
That is the whole payload. Your journal, your Why, your notes, your free text, your email, and your identity are never sent to the AI. The coach is architecturally prevented from reading your journal or Why content.
Every generated message passes through a deterministic server-side filter before it can be shown. If a message violates the app's language rules, a fixed pre-written fallback is shown instead. No unfiltered AI text ever reaches your screen. Generation is rate-limited per user.
AI output is generated content. It can be imperfect, generic, or occasionally miss the mark. It is not professional advice, not therapy, and not a clinical judgement about you. See the Medical and Clinical Disclaimer (/medical-disclaimer).