How to prepare birth chart information for an AI explanation

Prepare an AI chart explanation by sharing a clearly labeled, already calculated chart, preserving its settings and uncertainties, and asking one focused question. A cropped wheel without a legend or a birthday without a calculated chart leaves room for assumptions you may never notice.

Our guide to AI birth-chart explanations covers what an AI can and cannot do. This page is the practical handoff: what to copy, what to leave out and how to make the first response checkable.

Make a compact chart packet

Use the same source throughout the packet. Record the calculator’s name and the settings displayed with the chart. For a Western chart, include the zodiac and house system if shown, then copy the placements you want explained. For Saju, label the year, month, day and hour pillars and identify an unknown hour explicitly.

The TimePassages manual documents birth-data accuracy notes and house-system settings. Those details belong with its chart output; they should not disappear when you move the information into a chat.

You do not need to upload a birth certificate to explain a placement. Remove account details, unrelated names and document identifiers. If you are asking only about a chart table, share the relevant rows rather than a whole personal file. Check the receiving service’s data controls before sharing.

Use a handoff like this

The following is a template, not a calculated example:

Tradition: [Western natal astrology or Korean Saju]. Source: [calculator]. Settings: [as displayed]. Birth-time confidence: [known, approximate or unknown]. Chart entries: [paste labeled rows]. First repeat the entries you can read. Flag missing or ambiguous information. Then explain [one question] using only the supplied chart. Separate traditional interpretation from practical suggestions.

If the AI cannot read a character or degree, replace that portion with text. Ask it to stop at the missing information instead of inventing a complete chart.

Why the confirmation step matters

A September 2025 OpenAI evaluation reported a 26% error rate for gpt-5-thinking-mini on SimpleQA. OpenAI’s report concerns a historical factual benchmark, not astrology, Saju or today’s models. It is a reason to verify supplied facts rather than treating fluent wording as a successful chart transcription.

For example, if your source labels an hour unknown but the response describes an hour pillar, correct that before discussing its meaning. An interpretation built on an invented entry is not repaired by sounding plausible.

Keep the question small

Start with “Explain this labeled placement” or “Which supplied feature are you referring to?” Add another question once the first explanation matches the chart. Save the original packet so later replies can be compared against it.

For a fuller review after the explanation arrives, use our AI reading audit checklist. Getting the information into a chat and checking the resulting claims are separate steps.