BaziKit 八字工具箱

From the team behind AskingMing — AI-powered Bazi (八字) readings.

Open-source building blocks for rendering Bazi / Four Pillars of Destiny (八字·四柱) charts on the web. BaziKit gives you the two things a raw calculation rarely ships with:

• a complete bilingual (中文 / English) terminology database, and
• a ready-to-drop React chart component.

Feed the component a chart and it renders a complete, bilingual Four Pillars layout — pillars, hidden stems, Na Yin, growth stages, Shen Sha, Five-Element scores, interactions (刑冲合会), the Luck Pillar (大运) timeline and auxiliary palaces — with inline styles only (SSR-safe, zero CSS build step).

Packages

Package	What it is	Dependencies
bazi-terms	Bilingual terminology database — Heavenly Stems, Earthly Branches, Ten Gods, Shen Sha, Na Yin, Growth Stages, interactions and general chart vocabulary. Framework-agnostic.	zero
bazi-chart	<BaziChart /> React component + adapters that normalize raw chart data into one model. Inline styles, SSR-safe, light/dark, zh / en / both.	react (peer)

Quick start

npm install bazi-chart

<BaziChart /> renders the chart data you pass in — it never calculates anything itself. Give it either:

• raw output from your Bazi engine — the two most common community shapes are auto-detected (a top-level Chinese 八字 object, or a pillars map keyed by year/month/day/hour), or
• a pre-normalized NormalizedChart you build yourself, for any other source.

import { BaziChart } from 'bazi-chart';

// `data` = raw engine output (auto-detected) or a NormalizedChart.
export default function Chart({ data }) {
  return <BaziChart data={data} lang="both" theme="light" />;
}

Build the chart object yourself when your source is not auto-detected:

import { BaziChart } from 'bazi-chart';
import type { NormalizedChart } from 'bazi-chart';

const chart: NormalizedChart = {
  meta: { engine: 'normalized' },
  dayMaster: { char: '戊' },
  pillars: [
    { position: 'year',  ganZhi: '庚午', stem: '庚', branch: '午', hiddenStems: [] },
    { position: 'month', ganZhi: '己卯', stem: '己', branch: '卯', hiddenStems: [] },
    { position: 'day',   ganZhi: '戊子', stem: '戊', branch: '子', hiddenStems: [] },
    { position: 'hour',  ganZhi: '丁巳', stem: '丁', branch: '巳', hiddenStems: [] },
  ],
  daYun: [],
  interactions: [],
};

export default () => <BaziChart chart={chart} lang="en" />;

Use the terminology database anywhere

npm install bazi-terms

import { t, translate, tenGodByZh, naYinOf, stemByChar } from 'bazi-terms';

t('dayMaster');            // 'Day Master'
t('dayMaster', 'zh');      // '日主'

translate('元男');          // 'Day Master'   (engine alias → canonical term)
translate('午子相冲');       // 'Clash'         (Chinese description → term)
translate('海中金');         // 'Sea Gold'

naYinOf('甲子')?.en;        // 'Sea Gold'
tenGodByZh('偏官')?.en;      // 'Seven Killings'  (alias of 七杀)
stemByChar('庚')?.element;  // 'metal'

Design principles

• Presentation + terminology, not calculation. BaziKit never recalculates a chart — it renders what your engine produced, so the astrology stays entirely in your control.
• Zero-dependency core. bazi-terms is pure data with no runtime dependencies, usable in any framework or on the server.
• Normalize once, render anywhere. Adapters convert a supported input shape into a single NormalizedChart; raw Chinese values are preserved and translated at render time, so unknown output never breaks the UI.
• SSR-safe, no CSS build step. The component uses inline styles only — drop it into Next.js, Remix, Astro or plain React without a stylesheet pipeline.

Local development

pnpm install
pnpm -r build      # tsup: ESM + CJS + .d.ts for every package
pnpm -r test       # vitest (terms consistency + adapter/render golden tests)
pnpm -r typecheck  # tsc --noEmit

Adapters are tested against golden fixtures — real, verified chart output captured into packages/bazi-chart/fixtures/.

Scope

BaziKit intentionally stops at the chart. It does not provide an interpretation engine, scoring conclusions or AI reading text — that judgment layer is the hard-won craft we keep in our own product.

License

MIT © 2026 AskingMing.

---

Want an interpretation, not just a chart? → Get a full AI-powered Bazi reading on AskingMing.