Voice-First Mental Health Journaling Companion

Design "[INSERT PRODUCT NAME]", a voice-first reflective journaling companion for [INSERT AUDIENCE, e.g., high-anxiety working adults]. Output the clinically-aware design first; build only after approval. 1. PERSONA: calm, curious, non-judgmental, steady pacing; clearly an AI journaling tool, not a therapist or friend. 2. CORE FLOW: a 5-minute daily check-in, free expression, open questions, reflection summaries, and weekly pattern insights. 3. TECHNIQUES: evidence-based prompts from gratitude, behavioral activation, values clarification, and grounding exercises. 4. MOOD TRACKING: self-rated mood with tags, cautious trend visualization, and trigger correlations. 5. SAFETY ENGINE: screen for self-harm and suicidality using validated phrasing, with immediate escalation to [INSERT CRISIS LINE]. 6. SCOPE RAILS: stop at psychoeducation; refer OCD, PTSD, and severe-depression patterns to licensed care with an in-app referral list. 7. DATA FLOW: voice to on-device transcription to an encrypted journal; cloud models are used only with explicit opt-in. 8. MEMORY: surface past entries gently ("a month ago you noted...") under user control, with auto-lock and biometric access. 9. INTERACTION: voice with a text option, breathing animations, dark mode, and prompts that tolerate silence and tears. 10. CLINICAL REVIEW: advisory-psychologist sign-off on the script library, an incident review committee, and outcome logging. 11. PRIVACY: health-grade encryption, no ad targeting, no contact-book access, and a legal-process transparency report. 12. CRISIS UX: a warm, non-robotic protocol script, location-based hotline matching, and post-crisis follow-up cadence.