Skip to main content

Serenity

Our Ethical Commitments

Building for health carries profound responsibility. Serenity is committed to these five principles in everything we design, build, and maintain.

Privacy First

Health data is among the most sensitive. We design with privacy by default.

We do

  • Encrypt sensitive data with AES-256-GCM before storing
  • Minimise data collection to only what is necessary
  • Provide a clear, honest privacy policy
  • Allow full data export and deletion at any time

We don't

  • Store unnecessary data
  • Share data with third parties (beyond AI processing disclosed in Privacy settings)
  • Use dark patterns to discourage data deletion

Do No Harm

This app may reach vulnerable people. We build safeguards against misuse.

We do

  • Display crisis resources (988, Crisis Text Line) prominently on the AI page
  • Detect crisis keywords and prioritise emergency resources in AI responses
  • Include disclaimers that the AI is not a therapist or doctor
  • Recommend professional help for serious concerns

We don't

  • Provide medical diagnoses
  • Recommend specific medications or treatments
  • Replace professional mental health help
  • Ignore crisis signals in user messages

Inclusive Design

Mental health affects everyone differently. This app must be accessible to all.

We do

  • Follow WCAG 2.1 AA guidelines for accessibility
  • Provide aria-label on all interactive elements
  • Support keyboard navigation with visible focus rings
  • Include a skip-to-main-content link for screen reader users
  • Use aria-live regions for dynamic content updates

We don't

  • Assume abilities or technical literacy
  • Use colour alone to convey meaning
  • Create icon-only buttons without accessible names
  • Ignore screen reader compatibility

Transparency

Users should understand how this tool works, especially where AI is involved.

We do

  • Disclose that this app is powered by Google Gemini AI
  • Explain what data is sent to the AI and why
  • Document limitations clearly (not a therapist, not infallible)
  • Be honest about what the AI can and cannot do

We don't

  • Hide AI involvement
  • Overstate the effectiveness of AI wellness advice
  • Use manipulative or deceptive design patterns
  • Obscure data flows from users

Sustainable Impact

Build solutions that can grow and be maintained. Consider long-term viability.

We do

  • Write clear documentation (README, ETHICS.md, code comments)
  • Use maintainable, well-structured architecture
  • Plan for handoff and future maintainers
  • Keep dependencies up to date and minimal

We don't

  • Create unnecessary technical debt
  • Ignore security patching and dependency updates
  • Build features without documentation
  • Optimise only for short-term delivery

Our Commitment

These are not just words — they are implemented in the codebase. Serenity uses AES-256-GCM encryption for sensitive data, includes always-visible crisis resources, follows WCAG accessibility guidelines, discloses all AI usage, and is documented for future maintainers. We hold ourselves accountable to these standards.

Last reviewed: March 2026