English Français

Clinical Safety

Effective date: September 14, 2026

Mental Detox is a mental-wellness and self-reflection tool. It is not a medical device and is not subject to regulation as Software as a Medical Device (SaMD) under Health Canada guidance. For the full rationale, see section 1.3 of the Terms of Service.

This document describes the risks we have identified for users and the measures taken to reduce them.

1. Scope — Why the OCSA Does Not Apply

Mental Detox does not perform any function that falls within the scope of the ORCHA Clinical Safety Assessment (OCSA): it does not diagnose, treat, contain clinical calculators, transmit data to clinicians, or automate clinical decisions. The Service provides emotional journaling and evidence-based self-awareness support.

2. Identified Hazards

#	Hazard	Severity	Likelihood
H1	A user in crisis relies on the AI instead of emergency services	High	Low
H2	The AI provides an incorrect, harmful, or inappropriate response	Medium	Low
H3	A user develops emotional dependence on the AI assistant	Medium	Medium
H4	A minor uses the Service without supervision	Medium	Low
H5	A user interprets AI responses as a diagnosis or prescription	High	Low
H6	Loss of or unauthorized access to emotional journal entries	Medium	Low
H7	A user delays seeking professional help because AI support feels "sufficient"	Medium	Medium

3. Risk Controls

H1 — Crisis situation

A crisis protocol is built into all four AI modules with the highest priority.
Crisis helpline numbers are displayed in the footer on every page.
Section 1.2 of the Terms clearly defines who should not rely on the Service.
A callout block with 911 and 988 numbers appears in the Terms.

H2 — Incorrect AI response

All techniques are selected based on evidence (24 techniques documented in a dedicated report).
Section 1.1 of the Terms warns that AI responses may be inaccurate.
AI system prompts contain an explicit prohibition on diagnosing or prescribing treatment.
A one-technique-per-response principle reduces cognitive overload.

H3 — Emotional dependence

Section 1.1 of the Terms directly warns about the risk of emotional reliance on AI.
The AI does not simulate friendship, romance, or a therapeutic relationship.
System prompts include instructions to encourage human contact.

H4 — Minors

Minimum age is 16 (section 2 of the Terms).
Terms and Privacy acceptance checkboxes are required before registration.
The Service is not directed at children and does not collect children's data.

H5 — Perceived as diagnosis

Section 1 of the Terms states: "does not diagnose… does not prescribe treatment."
Section 1.3 of the Terms confirms: "not a medical device."
AI system prompts contain: "You are NOT a doctor, psychiatrist, or psychotherapist."

H6 — Data security

Data is stored in Firebase with encryption.
Authentication is handled through Firebase Auth.
Full details are available in the Privacy Policy.

H7 — Delayed professional help

The app is positioned as a tool "between sessions with a psychologist."
The AI directs users to a professional when signs of a serious condition are detected.
Section 1.2 of the Terms defines the Service's limitations.

4. Residual Risks

After the controls described above, residual risks are considered acceptable. The user remains responsible for:

contacting emergency services promptly when there is a threat to life;
following the recommendation to consult a qualified professional for serious conditions;
maintaining the security of their own device and account.

5. Methodology and Qualifications

A detailed description of the 24 evidence-based psychological techniques, the crisis protocol, technique-selection algorithms, and the full list of scientific references is available in the document "Report on Evidence-Based Psychological Techniques Used in the Mental Detox App" (version 1.0, April 2025).

The crisis protocol is based on:

Safe Messaging Guidelines (American Foundation for Suicide Prevention)
Motivational Interviewing (Miller & Rollnick, 2002)
DBT Crisis Survival Skills (Linehan, 1993)

6. Accessibility and Design Standards

Mental Detox is developed with reference to WCAG 2.1 AA, Apple Human Interface Guidelines, and Android Application Quality Guidelines. The Service supports dark mode, uses the device's system font that scales automatically according to the user's accessibility settings, and is built with a responsive layout for correct display across different screen sizes. We continue to improve the accessibility of the Service.

7. User Involvement and Feedback

Mental Detox was created by a founder who is a combat veteran with a background in nursing and psychology, and firsthand experience of stress, emotional exhaustion, and post-war adjustment. During development and testing, feedback was collected from individuals with their own lived experience of stress, anxiety, and emotional difficulties. Guidance on the clinical approach and tone of the Service was also provided by psychologists. This input influenced feature selection, the tone of AI responses, and the overall approach to user interaction. We continue to collect feedback to improve the Service.

8. Continuous Development and Support

Mental Detox is actively developed and maintained. Updates are released on an ongoing basis, informed by new knowledge, recommendations from psychologists, and user feedback. Feedback is accepted through support@mental-detox.app, as well as through reviews on the Apple App Store and Google Play.

Contact

If you have questions or concerns about clinical safety, contact:
support@mental-detox.app