What goes into an app?

Ingredients: features and patterns

A reader, prayer list, AI companion or subscription is a feature. A pattern describes the choices within it: how a streak handles a missed day, how an assistant names its role, or how a prayer becomes a handover to another person.

We link observed ingredients to reviewed screenshots and the relevant pattern articles. Captures show particular states, versions and dates. Publisher screenshots are separately labelled marketing material. Missing ingredients mean “not yet verified”, rather than “absent”.

The same ingredient can support different experiences. A reminder can be easy to control or difficult to escape; a paid plan can explain its terms early or reveal them late. The implementation, defaults, controls and audience belong in the assessment.

Exploitative, ethical, redemptive

Our approach adapts Praxis’s Redemptive Frame alongside our design principles. This is our adaptation, not Praxis certification or endorsement.

Exploitative concern

Evidence that a particular design takes advantage of a person: for example through coercion, deception, extraction or a misleading claim of spiritual authority. A concern needs evidence of the actual behaviour; a subscription or unpopular price alone does not establish it.

Ethical safeguard observed

A specific interaction respects agency, tells the truth, protects dignity or provides a clear boundary. This is a finding about the evidenced interaction, not a seal of approval for the entire app.

Redemptive practice

A concrete loss or exclusion is addressed through restoration, with cost or risk freely borne by the builder and a credible mechanism benefiting the people served. A promising screen can be a candidate while the cost, operation or outcomes remain unverified.

Eight questions, with room for unknowns

• Money and access: Can people understand the cost, enter freely and leave clearly?
• Attention and agency: Can they set a rhythm, pause and regain control?
• Truth and spiritual authority: Are sources, uncertainty and human or machine roles clear?
• Privacy and dignity: Are sensitive disclosures treated with care?
• Formation and relationships: Does the design make room for God, other people and life beyond the app?
• Care and inclusion: Who can participate, and who needs particular protection?
• Craft and repair: Does the product work, and can people recover when it fails?
• Stewardship and accountability: Who is answerable for the product and its effects?

We publish scoped observations and their limits. “Not assessed” is not a pass. We do not assign a total score: serious concerns should remain visible beside strengths, and missing evidence should not become a favourable grade.

Screenshots cannot establish backend data handling, workplace conduct, leadership intent or spiritual growth. Those questions need other evidence. Christian branding, charity status, free access and high store ratings do not by themselves establish redemptive practice.

What user response can tell us

App profiles keep Apple and Google Play samples separate. We show sample size, collection date, selection method and the requested language/storefront. Those requests do not identify a reviewer’s residence.

• Google Play: up to 50 newest written reviews per English/US and German/DE request, only after matching the Android product. The displayed counts include reviews within the 180 calendar days ending 25 October 2026. Samples under 20 are marked small; language streams can overlap and should not be added.
• Apple: excerpts returned by the public review page in its default order. These are selected, sometimes historical excerpts, not a recent or random sample. Empty results do not mean there are no reviews.
• Star counts: 4–5, 3 and 1–2 stars describe ratings among retrieved written reviews. They are not text sentiment, the store’s overall rating or an ethical score. High-star reviews can still contain substantial complaints.
• Illustrative reading: concise paraphrases of up to two selected excerpts per app. These suggest questions for investigation; they are not exhaustive thematic coding or verified findings about every complaint.

Doctrinal disagreement, objections to paying for religious content and reports of a broken payment flow are different claims. Review reports need context and corroboration. Personal comfort is not evidence of clinical benefit or measured spiritual formation.

General web discovery remains research material while we check identity, source authority and repeated or syndicated claims. Direct Google Search was unavailable during this pass; no “overall internet sentiment” score is published.

Store context: Apple ratings and reviews · Google Play ratings and reviews. Observable manipulation framework: FTC, Bringing Dark Patterns to Light; this reference does not constitute a legal finding about a named app.

Corrections and builder responses

Apps change. If an example is outdated, an identity is wrong or important context is missing, send a correction with the app, platform, version and supporting evidence. Builders can offer context and responses; payment does not change an assessment.

Profiles are a growing research collection. Coverage varies, and the published screenshots are selected examples rather than complete walkthroughs.