Product quality

Seeded examples help Leyra improve safely

Leyra uses synthetic logged-in examples to review paid journeys, personalisation, emails, audio-pack scripts and WhatsApp previews without exposing private user readings or memory.

Downloadable audit fixtures

Rendered examples for AI review, demos and UX checks

Each bundle includes a rendered HTML preview, SVG image, simple PDF, README and AI-review prompt. These are synthetic and safe to pass into audit threads.

Free new user

A first-time visitor who wants a gentle daily sign reading without committing to memory yet.

Signals: Profile basics partial · memory not active · no saved readings

Plus evening ritual

A Plus user returning at night for moon rhythm, saved history and a calmer closing prompt.

Signals: Email enabled · evening ritual due · profile basics ready

Pro with active memory

A Pro user with memory consent, check-ins, guidance style and reflection-profile signals active.

Signals: Memory active · guidance style saved · reflection profile medium confidence · tarot available

Professional reset

A busy professional using Leyra as a private reflective pause during the working week.

Signals: Work focus · practical tone · weekday reminders · journal available

Synthetic, not private

What gets audited

The aim is simple: make the logged-in product feel clearer, more personal and more trustworthy before more users arrive.

Paid-plan journeys

Free, Plus and Pro states can be reviewed without exposing a real user account.

Personalisation proof

Examples show how memory, guidance style and reflection-profile signals should change tone without becoming diagnostic.

Future channels

Email, audio and WhatsApp copy can be tested as scripts and previews before live delivery.

Privacy boundary

No private reading text is published

  • Seeded examples use fictional users and synthetic profile signals.
  • External AI audits should receive screenshots or synthetic exports, not private account data.
  • Real memories, check-ins, WhatsApp bodies and journal text remain private unless a user explicitly exports them.