Privacy and data handling

What a free cast stores

• An anonymous session identifier in a signed cookie, valid for 30 days, used only to apply the three-casts-per-day limit.
• The cast itself: the hexagram, the changed hexagram, the moving lines, the topic you selected, a coarse referrer group, and a timestamp.
• Interaction events such as cast started, cast completed, card downloaded, share opened, feedback submitted, and waitlist submitted.

Cast data and events are deleted automatically 90 days after they are created.

What is never stored

• Your IP address is not written to the database or the logs.
• Your question is never sent to us. The interface asks you to hold it in mind; it is not collected, transmitted, or logged.
• No email address is required. If you voluntarily join the waitlist, that address is stored separately from the anonymous cast events and removed on request.
• No advertising or analytics cookies are used.

Who processes it

• Cloudflare hosts the site and provides the network in front of it.
• Supabase stores the anonymous cast sessions, readings, and events.

Paid Bazi reports

Paid reports are ordered and delivered through Fiverr, whose own privacy terms also apply. To calculate a chart, the birth date, time, city, and the calendar type you supply are used. Only normalized chart facts are sent to the third-party model provider that drafts the prose; the raw birth record is not sent. Order intake data is kept in private storage for at most 30 days and then deleted. The audit record kept for support contains the order identifier, engine and model versions, timestamps, and a PDF hash, never your birth data or the report text.

Deletion

Write to support@cheninsight.com to have your waitlist email, a specific cast, or the data related to a paid order deleted. Anonymous casts cannot be linked back to a person, but a waitlist address and an order can both be removed.

Changes

If this policy changes in a way that affects what is stored, the change will be described here with the date it takes effect.