portable-text-serialization

作者: sanity-io

將 Portable Text 渲染並序列化為 React、Svelte、Vue、Astro、HTML、Markdown 及純文字。適用於在任何前端實作 Portable Text 渲染時使用…

npx skills add https://github.com/sanity-io/agent-toolkit --skill portable-text-serialization

Portable Text Serialization

Render Portable Text content across frameworks using the @portabletext/* library family. Each library follows the same component-mapping pattern: you provide a components object that maps PT node types to framework-specific renderers.

Portable Text Structure (Quick Reference)

PT is an array of blocks. Each block has _type, optional style, children (spans), markDefs, listItem, and level.

Root array
├── block (_type: "block")
│   ├── style: "normal" | "h1" | "h2" | "blockquote" | ...
│   ├── children: [span, span, ...]
│   │   └── span: { _type: "span", text: "...", marks: ["strong", "<markDefKey>"] }
│   ├── markDefs: [{ _key, _type: "link", href: "..." }, ...]
│   ├── listItem: "bullet" | "number" (optional)
│   └── level: 1, 2, 3... (optional, for nested lists)
├── custom block (_type: "image" | "code" | any custom type)
└── ...more blocks

Marks come in two forms:

  • Decorators: string values in marks[] like "strong", "em", "underline", "code"
  • Annotations: keys in marks[] referencing entries in markDefs[] (e.g., links, internal references)

Component Mapping Pattern (All Frameworks)

Every @portabletext/* library accepts a components object with these keys:

KeyRendersProps/Data
typesCustom block/inline types (image, code, CTA)value (the block data)
marksDecorators + annotationschildren + value (mark data)
blockBlock styles (h1, normal, blockquote)children
listList wrappers (ul, ol)children
listItemList itemschildren
hardBreakLine breaks within a block

Framework-Specific Rules

Read the rule file matching your framework:

  • React / Next.js: rules/react.md@portabletext/react or next-sanity
  • Svelte / SvelteKit: rules/svelte.md@portabletext/svelte
  • Vue / Nuxt: rules/vue.md@portabletext/vue
  • Astro: rules/astro.mdastro-portabletext
  • HTML (server-side): rules/html.md@portabletext/to-html
  • Markdown: rules/markdown.md@portabletext/markdown
  • Plain text extraction: rules/plain-text.md@portabletext/toolkit

Additional Community Serializers

These are listed on portabletext.org but don't have dedicated rule files:

TargetPackage
React Native@portabletext/react-native-portabletext
React PDF@portabletext/react-pdf-portabletext
Solidsolid-portabletext
Qwikportabletext-qwik
Shopify Liquidportable-text-to-liquid
PHPsanity-php (SanityBlockContent class)
Pythonportabletext-html
C# / .NETdotnet-portable-text
Dart / Flutterflutter_sanity_portable_text

Common Patterns (All Frameworks)

Custom Types Need Explicit Components

PT renderers only handle standard blocks by default. Custom types (image, code, callToAction, etc.) require explicit component mappings — they won't render otherwise.

Keep Components Object Stable

In React/Vue, define components outside the render function or memoize it. Recreating on every render causes unnecessary re-renders.

Handle Missing Components Gracefully

All libraries accept onMissingComponent to control behavior when encountering unknown types:

  • false — suppress warnings
  • Custom function — log or report

Querying PT with GROQ

Always expand references inside custom blocks:

body[]{
  ...,
  _type == "image" => {
    ...,
    asset->
  },
  markDefs[]{
    ...,
    _type == "internalLink" => {
      ...,
      "slug": @.reference->slug.current
    }
  }
}

來自 sanity-io 的更多技能

sanity-migration
sanity-io
規劃、執行並審查從其他內容管理系統及內容平台遷移至 Sanity 的作業。適用於從 AEM、Adobe Experience Manager、Contentful、Strapi、Webflow、WordPress、Payload、Drupal、Markdown/MDX/frontmatter 檔案、WXR/XML 匯出、CMS API、資料庫匯出、靜態 HTML 進行遷移或平台轉換,或設計資料擷取、轉換、Portable Text 轉換、資產遷移、重新導向、驗證及切換流程時使用。
officialdevelopmentdatabase
create-agent-with-sanity-context
sanity-io
透過 Agent Context 建立對 Sanity 內容有結構化存取權限的 AI 代理。適用於設定由 Sanity 驅動的聊天機器人、將 AI 助手連接到 Sanity…
official
dial-your-context
sanity-io
互動式工作階段,用於為 Sanity Agent Context MCP 建立「指令」欄位內容。每當使用者提及調整代理上下文、改善…時,請使用此技能。
official
optimize-agent-prompt
sanity-io
透過引導式對話調整您的 Sanity Agent Context 代理。將探索數據轉換為可投入生產的指令,並精心設計系統提示…
official
shape-your-agent
sanity-io
互動式工作階段,用於為由 Sanity Agent Context MCP 驅動的 AI 代理設計系統提示。當使用者想要定義代理人格時,請使用此技能…
official
content-experimentation-best-practices
sanity-io
結構化指引,涵蓋設計、執行與分析內容實驗以提升轉換率與參與度。內容包括假設框架、指標選擇、樣本數計算,以及A/B測試與多變量實驗中的統計顯著性檢定。提供關於p值、信賴區間、統計檢定力分析及貝氏方法的詳細資源,用於解讀實驗結果。同時包含CMS整合模式,以便在欄位層級管理變體,並連接外部系統。
official
content-modeling-best-practices
sanity-io
結構化內容建模指南,涵蓋綱要設計、可重用性與多渠道發布。核心原則包括:將內容視為數據而非頁面、維護單一事實來源、為未來渠道設計,以及優化編輯工作流程。提供關於引用與嵌入物件、關注點分離及內容重用模式的決策框架。包含針對扁平、階層與分面分類法的分類學指引。適用於...
official
portable-text-conversion
sanity-io
將 HTML 和 Markdown 內容轉換為適用於 Sanity 的可攜式文字區塊。用於從舊版 CMS 遷移內容、將 HTML 或 Markdown 匯入 Sanity 等情境。
official