Editorial and AI policy

What the app presents

Card names, short themes, prompts, reversal notes, and method guidance ship in the application bundle. For a given card and method, visitors receive the same stored guidance; no model interprets your intention or reflection at runtime.

Editorial language should frame tarot as a reflective practice, distinguish sourced history from project-authored interpretation, and avoid claims of prediction, diagnosis, certainty, or professional authority.

AI assistance

AI tools assisted with drafting this initial trust-page foundation. AI output is not treated as a source. Claims on these pages were checked against the application code, tests, deployment configuration, and stored provenance records.

Future AI-assisted changes must remain reviewable in source control, identify the evidence for factual and rights claims, and receive maintainer review before publication. Material AI assistance should be disclosed on the affected page or in the project changelog.

Visitor content

The app does not send intentions, reflections, tags, or imported journals to an AI service. It has no remote interpretation API. If that behavior changes, the privacy page and this policy need review before the feature is published.