How to Prompt Xingque AI: Chart Notes, Analysis Goals, and Better AI Questions
A practical guide for users who have configured the Xingque AI Key and want better Horosa AI analysis through clearer chart context, goals, and output formats.
Turn vague requests into concrete tasks
After setting up the AI Key, many users start with a request like “please read this chart.” That is too broad, so the model often returns generic descriptions. A better prompt names the subject, the analysis scope, and the kind of result you want.
For example: “Based on the current chart, identify the three most important structures, then explain the possible evidence, what information needs verification, and where uncertainty remains.” This gives AI a reviewable analysis frame.
Give AI three things: chart facts, goal, and format
The first part is chart facts: current chart, selected technique, time and location details, and the conditions you have already verified. The second part is the goal, such as personality structure, career clues, relationship themes, timing triggers, or reference organization.
The third part is output format. Ask for “observation / evidence / possible reading / information to verify,” or ask for an outline before detailed analysis. Clear formatting makes the output easier to save into case notes.
Do not ask AI to be the final judge
Xingque AI is best used as an organizer and analysis assistant, not as the final authority. In astrology and Chinese metaphysics workflows, many judgments depend on background context, practitioner experience, and later verification.
A safer prompt is: “List possible interpretations and show which chart details each one depends on. If context is insufficient, list the questions I should answer next.” This reduces overconfident output.
A reusable prompt template
Start with this template: “Please produce a structured analysis based on the current chart. My goal is [insert goal]. List 3-5 main observations. For each one, include chart evidence, possible interpretation, information to verify, and follow-up questions. Avoid absolute conclusions and mark uncertain areas clearly.”
For long-term case work, keep AI output and your final judgment separate. Let AI draft observations and questions, then use your own review to edit, confirm, and refine the method.
