| name | Mind Map HTML Output |
| description | Generate a self-contained, interactive HTML mind map file powered by the Mind Elixir core. Use when a user wants to turn outlines, markdown, notes, or structured text into a standalone .html file that renders a editable mind map in any browser (a "canvas-capable" client) — no build step required. |
Mind Map HTML Output
Turn any hierarchical content (bullet outline, markdown headings, meeting notes, a topic tree, structured JSON) into a single standalone .html file that renders an interactive mind map using the Mind Elixir core loaded from a CDN. The file opens directly in any browser — it is the "canvas client". No npm install, no bundler.
This is the right skill when the user says things like "把这段内容做成思维导图", "export a mind map as html", "生成一个可以直接打开的思维导图网页", or "用 mind-elixir 内核输出 html 思维导图".
When to use
- The deliverable is a file the user can open/share, not a component inside an existing app.
- The user wants zero-setup output (double-click → mind map).
- Input is prose/outline/markdown that must first be structured into a tree.
If the user instead wants to embed Mind Elixir into a React/Vue project, use the Integrate Mind Elixir skill. For image-only export, use Export Mind Map as Image.
Workflow
- Collect / parse the content. Accept free text, markdown, or a tree. Reduce it to a single-rooted hierarchy.
- Build the
MindElixirData JSON (see below). Prefer JSON when you need styling, icons, tags, links, or images. For a plain nested-list outline you may instead use plaintextConverter (optional, see end).
- Produce the HTML file from the template
assets/template.html: replace __TITLE__ with a title and __MINDMAP_DATA__ with the pretty-printed MindElixirData JSON object. Keep everything else intact.
- Deliver the
.html file via present_files so the user can open it in the built-in preview or a browser. Tell them the toolbar buttons (PNG / SVG / 保存数据 / 导出 HTML) let them re-export or snapshot edits.
MindElixirData shape
{
nodeData: {
topic: '中心主题',
id: 'root',
expanded: true,
children: [
{
topic: '分支 1',
id: 'b1',
expanded: true,
children: [
{ topic: '子节点 1-1', id: 'b1-1' },
{ topic: '子节点 1-2', id: 'b1-2' },
],
},
{ topic: '分支 2', id: 'b2' },
],
},
direction: 1,
theme: {
name: 'Latte',
palette: ['#dd7878', '#ea76cb', '#8839ef', '#40a02b', '#209fb5', '#1e66f5'],
cssVar: { '--main-color': , : , : , : },
},
}
Render options (set on new MindElixir(options)): el, direction: MindElixir.LEFT | MindElixir.RIGHT, toolBar, contextMenu, keypress, and optional markdown: (text) => parser(text) if topics contain markdown/rich text.
CDN notes (important)
- Pin to the major version so the API stays compatible:
https://cdn.jsdelivr.net/npm/mind-elixir@5/dist/MindElixir.js
https://cdn.jsdelivr.net/npm/mind-elixir@5/dist/MindElixir.css
- The CSS file is
MindElixir.css (not style.css) on the CDN — match the package's dist output.
- PNG/SVG export use Mind Elixir's built-in
mind.exportPng() / mind.exportSvg() — no @zumer/snapdom dependency needed for the standalone file.
Optional: convert a plain outline with plaintextConverter
If the input is already an indentation/markdown bullet list, you can convert it without hand-writing JSON. In a page/script with the module available:
import { plaintextToMindElixir } from 'https://cdn.jsdelivr.net/npm/mind-elixir@5/dist/PlaintextConverter.js'
const data = plaintextToMindElixir(`- 中心主题
- 分支 1
- 子节点
- 分支 2`)
Then pass data into the template's __MINDMAP_DATA__ slot. Plaintext format supports [^id] anchors, {color:...} node styling, } summaries, and > arrows — see the Streaming Mindmap skill for the full grammar.
Tips & pitfalls
- Always give the root and every node a unique, stable
id; regenerate ids deterministically from content paths to avoid the whole graph re-rendering on refresh.
- Keep topics short; long text hurts layout. Put detail in child nodes.
- The standalone file needs internet access on first open (CDN). If the user needs a fully offline file, download
MindElixir.js + MindElixir.css and rewrite the two CDN URLs to relative/local paths.
exportPng() is async and may return null if capture fails — the template guards against that.