Safety and privacy
Clear boundaries are part of the product.
You deserve to know what AI is doing, who can see their words, and what happens when writing suggests immediate danger.
AI is a tool
AI helps produce a brief acknowledgment, one follow-up question, a tiny-step idea, and weekly reflection. It is not a person, friend, therapist, or emergency service. Zevren does not use journal content to train models.
Account visibility in the existing beta
The verified adult who owns the account can read all journal entries, answers, tiny steps, and AI outputs. Students acknowledge this before their first reflection.
Urgent safety path
If automated checks classify writing as involving immediate danger or self-harm intent, normal coaching is suppressed. The student sees fixed guidance to go to a trusted adult now and to use 988 or 911 when appropriate. The account owner receives an alert with no journal excerpt.
Protected content
Journal text, AI turns, tiny steps, weekly narratives, and growth summaries are encrypted before database storage. Raw content is excluded from email, analytics, URLs, queue messages, and application logs.
Retention and control
Account owners can export their account data, delete individual entries immediately, or close the account. Product access is revoked when closure begins, with product data scheduled for hard deletion within 30 days, subject to documented backup expiration.
Small steps. Real growth.
Your next chapter starts with a small step.
Zevren is in private beta. Full launch is expected at the end of 2026.