AI safety and spiritual boundaries

• certainty about Allah's intent, the unseen, or the future;
• communication from or the condition of a deceased person;
• a medical, mental-health, legal, or religious ruling;
• that a dream proves guilt, infidelity, danger, or another person's motives;
• that an AI draft has received human review when it has not.

Drafts are constrained to reviewed source records and checked after generation. Missing sources, retrieval failures, invalid structure, invented citations, and quotation mismatches fail closed: no draft is displayed and no interpretation credit is consumed.

The service checks user text for indicators of imminent self-harm or harm. When detected, interpretation is paused and crisis resources are shown. This is a protective routing measure, not diagnosis or monitoring by a clinician. Tabeerly is not an emergency service; users in immediate danger should contact local emergency services.

Current automated safety screening and support content are validated in English only, and the listed national resources are U.S.-specific. People outside the U.S. should use local emergency or crisis services. Tabeerly does not claim multilingual safety coverage.

Dream text is sensitive content. It is encrypted at rest and in transit, but it is not end-to-end encrypted: Tabeerly systems and the configured AI provider must process it, and an assigned reviewer may access it during review. Dream text must not be sent to routine analytics events or written to ordinary application logs.

Users can stop before submitting, omit optional relationship context, export or delete account data, and revoke private gift links. Report a harmful response, unsafe claim, privacy issue, or source error to [email protected]. Include the interpretation ID, not the full dream text, unless that detail is necessary to investigate.