LLamjinLab
DestinyWords 2823Read time≈ 8 min

AI Metaphysics Project Collection: From Bazi and Ziwei to Feng Shui, Tarot, and Liu Yao

A curated set of open-source projects that combine traditional metaphysical systems, classical-text knowledge bases, Claude Skill, MCP, and AI interpretation, suitable for learning rule engineering, knowledge-base design, and cultural AI applications.

MetaphysicsBaziZiwei Dou ShuDestiny ChartsFeng ShuiTarotLiu YaoRAGKnowledge BaseClaude SkillMCPAI InterpretationDigitizing Traditional Culture

I. Introduction: What Is Actually Worth Looking At in AI Metaphysics Projects

Recently I have seen many projects that combine traditional metaphysical systems with AI, including Bazi, Ziwei Dou Shu, Qimen Dunjia, Feng Shui, Tarot, Liu Yao, Plum Blossom Numerology, Buddhist-scripture Q&A, and related areas.

These projects share a common trait: they do not rely on the large model to improvise freely. Instead, they integrate horoscope calculation workflows, classical text passages, rule-based judgments, card-meaning systems, or knowledge-base retrieval into AI conversations so that answers have more grounding and are easier to audit.

In other words, what is truly worth learning from them is not whether "AI can predict accurately," but how these projects decompose traditional knowledge into data, rules, processes, and tools. For AI Agents, Claude Skill, MCP, RAG, and knowledge-base engineering, such projects are actually excellent practical material.

Below are several representative projects. They can be used as cyber-relaxation toys, or as references for learning digital traditional culture, rule engines, knowledge-base design, and AI-assisted interpretation.

Reminder: The projects below are suitable only for technical learning, cultural exploration, and entertainment. Do not treat any result as a basis for real-life decisions. Your fate is still yours to decide; DYOR.

II. MetaphysicsQuant: Traditional Metaphysical Knowledge Base, Rule Engine, and Desktop Analysis System

This is my own project and the most complete engineering example in this list.

It is not a lightweight calculator that only displays Four Pillars, star charts, or hexagram charts. It attempts to turn the harder-to-engineer parts of traditional Chinese metaphysics into executable software systems, including calendar conversion, Heavenly Stems and Earthly Branch derivation, deity and pattern rules, classical text retrieval, topic-specific reports, AI second-pass interpretation, an SQLite case library, Markdown/TXT export, and a desktop workflow.

The current mainline is a Python-based desktop analysis system. Its core modules cover Bazi, Ziwei Dou Shu, I Ching divination, Liu Yao, auspicious-date selection and compatibility matching, name analysis, Western numerology, Tieban Shenshu, and Shaozi Shenshu. Compared with a single-purpose Skill, it is more like a "knowledge base + rule engine + case management + report generation" platform for the metaphysical domain.

The Bazi module is the heaviest part. Its focus is not just calculating the year, month, day, and hour pillars, but also building structured reports around the original chart, big luck cycles, yearly/lunar monthly flows, taiyuan, life palace and body palace, Seven-Pillar thematic interpretation, branch linking and fortune transfer, deity-patterns, classical passages, and AI interpretation entry points.

The Ziwei module centers on the Twelve Palaces, main stars, auxiliary stars, miscellaneous stars, four transformations, star brightness, major cycles, annual/monthly flows, and chart rendering. The I Ching and Liu Yao modules support time-based divination, number-based divination, original/changed hexagrams, moving lines, six relatives, six spirits, shiying/yingshen dynamics, na-jia, and related outputs. The date-selection module organizes rules by specific scenarios such as marriage, job seeking, opening a business, wealth, travel, medical visits, moving in, study, and childbirth.

This project is especially good for learning three things: first, how to break traditional rules into functions and JSON data; second, how to combine classical knowledge, chart outputs, and AI interpretation; third, how to turn a passion project into a long-running system with UI, database, report export, and case management.

Suitable for people who want to learn Python desktop systems, metaphysical rule engines, traditional knowledge bases, SQLite case management, AI-assisted interpretation, and report-export workflows.

III. Ziwei Knows: Ziwei Dou Shu Charting and AI Deep Interpretation Tool Based on iztro

Ziwei Knows is an open-source Ziwei Dou Shu charting and AI deep-interpretation tool with self-hosting support. Its positioning is not merely displaying charts; it combines Ziwei chart generation, AI interpretation, annual fortune analysis, composite relationship charts, and visual life trends, making it closer to a directly usable Web product.

The project uses the iztro library for Zhongzhou Pai Anxing-style star placement, with a frontend built using React, TypeScript, and Vite. Compared with traditional chart tools, it emphasizes interaction and productization: users can enter birth details to generate a Ziwei chart, then inspect AI interpretation, yearly trend analysis, partner compatibility charts, and modern visualizations such as a "life K-line."

A practical advantage is its relatively open model choice. The project documentation notes support for Kimi, Gemini, Claude, DeepSeek, and other model providers, making it easy for users to switch according to cost, quality, and availability. Deployment is also friendly, with options for self-hosting on Vercel or Cloudflare, making it a good reference for anyone quickly building a Ziwei AI application.

From a learning perspective, Ziwei Knows is a good case for the full chain of "traditional chart engine + modern frontend + multi-model AI interpretation + cloud deployment." If you want to make a metaphysical tool as a website instead of a local script or desktop app, this is a highly valuable open-source reference.

Suitable for people who want to learn Ziwei web applications, the iztro charting library, React + TypeScript frontends, multi-model AI integration, and self-deploying on Vercel/Cloudflare.

IV. Numerologist Skills: Putting a Cyber Spiritual Specialist Through a Safety Lock

The core idea of this project is to reduce LLM hallucinations in metaphysical analysis. A major weakness of traditional metaphysics is this: if the charting steps are wrong, fluent explanation afterward is meaningless. So it turns the core flows of Qimen Dunjia, Ziwei Dou Shu, Bazi, and related disciplines into fixed steps so AI is no longer left to unrestricted improvisation.

The project description emphasizes that it is an engineering framework for reducing hallucinations in Chinese astrology scenarios. In other words, it is not just a “you are a master” prompt; it turns deity selection, pattern analysis, chart judgment for Qimen, plus Ziwei and Bazi calculation and interpretation chains into more auditable, reusable, and extensible Skill modules.

From a learning perspective, it is a good case for observing how traditional rules are converted into AI-executable workflows. If you want to build your own metaphysical Agent, the most useful part is not its copywriting but how it turns vague analysis order into a clear chain: calculate first, validate next, interpret afterward, and only then provide advice.

Suitable for people interested in Qimen Dunjia, Ziwei Dou Shu, Bazi, low-hallucination Skill design, and metaphysical Agent workflows.

V. Cyber Divination / Bazi Skill: Classical Text-Based Four Pillars Analysis Tool

This is a Claude Code Skill specifically for Bazi analysis. Through interactive conversation it collects name, birthday, birth hour, gender, birthplace, and other details, then automatically generates a Four Pillars chart and analyzes day-master strength, Ten Gods relationships, five-element balance, chart structures, major luck cycles, annual luck, career, relationships, and health.

Its strength is that it combines Bazi analysis with classic astrology texts rather than relying on spontaneous model writing. The project description notes that it references 《穷通宝鉴》《三命通会》《滴天髓》《渊海子平》《子平真诠》 and related materials, which keeps the logic relatively stable across multi-turn follow-ups and enables continuation in traditional terminology.

This project is suitable for beginners to understand how a Bazi Skill should collect input. Bazi analysis is not enough with just a birthday; it also involves lunar/solar conversion, birth hour, gender, birthplace, and age-commencement details. The clearer the information collection flow, the less likely the downstream interpretation will drift.

Suitable for people who want to learn Claude Skill design, Bazi charting flow, interpretation of metaphysical terminology, and conversational information collection.

VI. Yue Lao Marriage-Fate Calculation Skill: A Modular Tool for Relationship Questions

This project focuses specifically on marriage and relationship analysis. Unlike broad metaphysical tools, its usage context is clearer: users are often not looking for a full chart, but for multi-angle guidance on dating, compatibility, relationship development, and romantic decision-making.

The project integrates Bazi compatibility matching, zodiac pairing, Ziwei spouse palace interpretation, divination for relationships, peach-blossom fortune analysis, and red-thread calculations into a more relationship-consultation-style Skill. Its value lies not in delivering absolute conclusions, but in organizing relationship-judgment approaches across traditions into one unified entry point.

From a product-design perspective, this type of Skill has an obvious advantage: the scope is narrowly vertical. Users do not need to understand complex metaphysical taxonomies; they ask a relationship question, and the system decides whether to invoke compatibility matching, zodiac pairing, divination, or other modules for auxiliary analysis.

Suitable for people who want to build vertical relationship Agents, marriage-fate calculators, and compatibility-analysis workflows.

VII. Master-skill: AI Learning Assistant Based on Buddhist Classics

This project is more focused on cultural-text learning than divination. It combines Buddhist canonical texts, master personas, and RAG retrieval so users can learn Buddhist content through conversation.

According to the project description, users can select among several masters from Chinese, Tibetan, and Theravada traditions, and responses are linked to textual evidence retrieval with citation of classical sources. It functions more like a "Buddhist learning assistant" or "digital humanities knowledge base," useful for understanding classics, terminology, sectarian thought, and different teachers’ explanatory styles.

Its key takeaway is that AI traditional-culture projects need not be limited to divination. They can support text study, retrieval, persona-style explanation, and cross-tradition comparison as well. For those exploring RAG, knowledge-base design, AI persona prompting, and source validation of classical references, it is a valuable reference.

Note that Buddhist AI learning tools are best used as reading aids and entry points, not as a replacement for formal study, lineage guidance, or rigorous academic research. The project itself includes a similar caution: content is generated from historical Buddhist sources and is for reference learning only.

Suitable for people who want to study Buddhist text Q&A, RAG retrieval, AI persona design, and digital humanities projects.

VIII. I Ching Feng Shui / Residential Analysis Skill: Turning Xuan Kong, Eight Mansions, and Auspicious-Date Practice into a Conversational Advisor

This project is positioned as a traditional Feng Shui advisor Skill, centered on living environment, house orientation, move-in timing, destiny trigrams, flying-star charts, Eight Mansions annual analysis, date selection, and common external-sha assessment.

A notable part is that the project explicitly emphasizes "Luantou is the form, Liqi is the usage" (Luantou is the essence; Liqi is the application). In other words, it does not rely on a single abstract direction; it tries to place form, qi governance, flying stars, Eight Mansions, and date selection into one coherent conversational framework.

From an engineering perspective, the challenge of a Feng Shui Skill is non-standardized input. Users may know the house faces south, the desk is near a window, and a road faces the entrance, but cannot provide exact orientation, move-in time, or floor plan. So this type of project is most valuable in how it iteratively gathers required details and translates Feng Shui terms into actionable layout suggestions users can understand.

Suitable for people who want to study Feng Shui knowledge engineering, space-oriented information gathering, Xuan Kong Flying Stars, Ba Zhai Mirroring, and date-selection rules.

IX. Tarot Reading Skill: A System of 78 Card Meanings and Spread Relations

This is an AI Tarot divination Agent Skill built around the 78 tarot cards, common spreads, inter-card relation systems, and random draw scripts. Compared to Eastern metaphysical projects, Tarot’s structure is clearer: deck, upright/reversed states, spread positions, question types, and card-to-card relations are easier to organize into a knowledge base.

The project description notes support for full 78-card meanings, six spread types, relational theory among cards, and true-random draw scripts. That means it does not simply let the model roll and speak; it first performs card selection and then interprets according to spread position and meaning framework.

This project is useful for observing the design pattern of "randomness + structured interpretation." In daily use, Tarot can function as multi-perspective reflection tool; for developers, it is a strong introductory Skill example because card meanings, spread structures, and interpretation rules are relatively easy to modularize.

Suitable for people who want to learn Tarot Agents, random card draws, spread structures, and lightweight Skill development.

X. Taibu: AI Divination Tool and MCP / Skills Project Covering Multiple Metaphysical Systems

Taibu is a broad-scope AI metaphysics project. Its description says it includes Bazi, Ziwei Dou Shu, Liu Yao, Plum Blossom Numerology, Qimen Dunjia, Da Liu Ren, Xiao Liu Ren, astrology, Taiyi Shenshu, Tarot, MBTI, face/hand reading, compatibility matching, daily/monthly fortune, and Zhou Gong dream interpretation, and supports MCP services and Skills.

If earlier projects are mostly single-direction or few-direction Skills, Taibu resembles a comprehensive metaphysics platform. Its emphasis is not depth in one module but unifying several common metaphysical entry points into one AI tooling ecosystem.

From a learning perspective, it is good for observing how MCP, Skills, web tools, and multi-module metaphysical entry points can be composed. Especially if you want to build a "unified-entry AI metaphysics platform," its overall structure is highly instructive.

Taibu also uses a mixed license model: some core packages are MIT, while other web, server, deployment, and runtime code is AGPL-3.0-only. That is usually fine for learning and personal use, but if you plan to redevelop and offer it as an online service, you should read the licensing terms carefully.

Suitable for people studying MCP, comprehensive AI metaphysics platforms, multi-module integration, and productized web deployment.

XI. Additional Observations: The Core Here Is Not Mysticism, but Engineering

Looking at these projects together, AI metaphysics projects tend to follow several pathways:

1. Charting-path: first compute charts, hexagrams, or spreads accurately, then let AI interpret them. 2. Skill-path: encode knowledge, process, and boundaries into SKILL.md so AI executes fixed steps. 3. RAG-path: turn classics, scriptures, and terminology explanations into searchable knowledge bases to reduce fabrication. 4. Web-product path: integrate charting, model selection, reporting, visualization, and deployment into a complete application. 5. MCP-tool path: expose metaphysical capabilities as external tools callable by Agents.

This also points to a larger issue: traditional culture AI projects are not just "prompt engineering." Long-maintained projects usually require data structures, rule decomposition, input validation, output templates, source references, test cases, and user boundaries.

This is especially true for metaphysics projects. Because they involve many terms, schools, and rules, without structured constraints a large model can sound fluent while its internal logic remains unreliable. The more complex the system, the less it can rely on a single claim like "you are a divination master."

XII. How to Study These Projects

If you only want to try them out, start with low-friction projects like Bazi, Tarot, and relationship-fate. If you want to learn implementation engineering, prioritize MetaphysicsQuant, Ziwei Knows, Taibu, Numerologist Skills, and the Feng Shui Skill. If you want to study knowledge bases and RAG, focus on Master-skill-type classics Q&A projects.

The learning order I recommend:

1. Start with the Tarot Skill to understand the basic form of knowledge base + random draw + structured interpretation. 2. Then study the Bazi Skill to learn how to collect necessary information through multi-turn dialogue. 3. Next, move to more complex-flow Skills like Qimen, Ziwei, and Feng Shui to understand how to reduce AI hallucinations. 4. Finally, study comprehensive projects like MetaphysicsQuant, Ziwei Knows, and Taibu to learn how to integrate multiple modules, frontend pages, databases, report export, AI interpretation, and cloud deployment into a complete system.

The real value in these projects is not whether they "predict accurately." It is how they turn traditional knowledge into data, rules, workflows, and tools. For AI Agent development, this is more valuable than writing prompts alone.

If you want to verify the chart cases mentioned in this article, you can use the site’s Bazi charting tool or Ziwei Dou Shu charting tool.

FAQ

How reliable are these AI metaphysics projects?

The quality varies significantly. Strong projects encode traditions accurately and provide clear calculation logic; weaker ones rely heavily on vague phrasing and generic scripts. It is recommended to validate outputs against classical sources or mature tools.

Which tools are more mature?

The site provides Bazi charting, Ziwei Dou Shu charting, and I Ching divination tools, whose logic is transparent and can be cross-checked against traditional methods.

Can AI really "tell fate"?

AI can handle rule-based and pattern-based computation within metaphysical systems, but "fortune interpretation" generally requires large case experience, judgment, and contextual background; AI still has clear limitations in this area.

How can I evaluate the quality of an AI metaphysical project?

Key indicators: whether chart outputs are accurate (cross-check with mature tools), whether interpretation logic is source-backed, and whether there are signs of over-marketing or manipulative persuasion.

References

Share

Share this article