html-deck（HTML deck 搜索）

HTML Slide Deck to PDF Workflow

Build polished, presentation-ready HTML slide decks and convert them to razor-sharp PDFs using headless Chrome.

When to use

• Creating a new slide deck from program documentation, notes, or outlines.
• Converting rough content into a professional presentation.
• Generating a PDF version of an HTML deck with exact page dimensions.
• Building decks that include app screenshots or product visuals.

Progress checklist

Deck Build Progress:
- [ ] Phase 1: Slide skeleton (HTML + CSS foundation)
- [ ] Phase 2: Content population (slides from source material)
- [ ] Phase 3: Visual polish (spacing, typography, layout fixes)
- [ ] Phase 4: Print-safe CSS
- [ ] Phase 5: PDF conversion (use bundled script; do not reimplement)
- [ ] Phase 6: Verification (screenshot comparison)

Phase 1 — Slide skeleton

Use the Domo theme — MANDATORY FIRST STEP

1. Read references/domo-theme.md — it contains a complete, copy-paste-ready HTML skeleton that produces slides matching the Domo template exactly.
2. Read references/assets/logo-data-uri.txt — it contains the actual Domo logo as a base64 PNG data URI. Replace every LOGO_DATA_URI placeholder in the skeleton with this value.
3. The logo PNG is also at references/assets/domologo.png.

Do not skip this step. Copy the complete skeleton from the theme file, then build slides into it. Every content slide MUST have:

• Header: <div class="slide-header"> with <h1>, .subtitle, and .header-line (thin gray separator — required)
• Content: <div class="slide-content has-bg"> — always use has-bg for the gray #DDE5ED background (this is the Domo look; content is vertically centered automatically)
• Footer: <div class="slide-footer"> with .confidential (left), .footer-line (center), and .domo-badge with <img src="LOGO_DATA_URI"> (bottom-right — required)

Dimensions

Standard 16:9 widescreen: --slide-w: 1024px; --slide-h: 576px;

Content area positioning

The has-bg class handles positioning automatically (position:absolute; top:90px; bottom:52px) with vertical centering (display:flex; justify-content:center). Do not override these with inline styles.

Phase 2 — Content population

Slide type patterns

Table slide — Use a styled <table> with labeled column headers. Keep cell padding tight (6px 10px) on dense slides.

Flow/pipeline slide — Horizontal flex row of boxes with arrow separators (▶). Use colored left borders or top borders to differentiate phases.

Card grid — Flex row of equal-width cards. Use flex:1 on each card, consistent padding, and subtle background colors (#F6F8FA).

Architecture diagram — Nested flex containers. Label each layer (e.g. "INTELLIGENCE LAYER", "STORAGE", "COMPUTE") with colored card groups.

Screenshot showcase — One hero image centered, or two side-by-side with max-width:48%. Apply border-radius:8px for polish. No box-shadow (removed by print CSS anyway).

Info boxes — Paired boxes below main content. Use display:flex; gap:32px with flex:1 on each.

Typography scale

Color conventions

Use CSS variables for consistency. Define brand colors, accent colors, and text shades in :root.

Phase 3 — Visual polish

Spacing rules

1. Content must never touch the footer. The bottom:52px on .slide-content creates a 52px exclusion zone. If content still overlaps, reduce margin-top, padding, gap, font-size, or line-height on the offending elements.
2. Header–content gap. The top value on .slide-content controls clearance below the header. Typical: 88–100px. Increase if subtitle text collides with content.
3. Dense slides. When a slide has both a table and info boxes, use compact values:

- Table cell padding: 4px 10px - Info box padding: 5px 12px - Gaps between cards: 3–5px - Font size: step down 1px from defaults

1. justify-content:center caution. Vertical centering can push tall content past bottom. Remove it on content-heavy slides and let content flow from top.

Image handling

• URL-encode spaces in src paths: Sample%20Assets/file.png.
• Avoid special Unicode characters in filenames. Copy to clean filenames if needed.
• Use max-width:100%; max-height:100% with object-fit:contain to auto-size within the content area.

Phase 4 — Print-safe CSS

PDF rendering through Chromium's print pipeline treats box-shadow differently than screen rendering. Shadows appear as heavy dark blocks in PDF output.

Required rule (already in the skeleton above):

@media print {
  *, *::before, *::after {
    box-shadow: none !important;
  }
}

This strips all shadows for print only. The HTML browser view retains its shadows.

Additional print considerations:

• page-break-after: always on .slide — one slide per page.
• page-break-inside: avoid — prevent mid-slide splits.
• background: white on body — no gray viewer background.

Phase 5 — PDF converter

Do not rewrite the converter from scratch. This skill ships a maintained script: references/convert-to-pdf.js. Copy it into the user’s project or run it by path from the skill directory after install.

npm install puppeteer-core
node convert-to-pdf.js path/to/deck.html path/to/deck.pdf
# optional second arg omitted => writes next to the HTML with .pdf extension

The script uses puppeteer-core with the system Chrome binary (not bundled Chromium), matching the approach below.

Key settings (implemented in the script; change only if you have a deliberate reason):

• deviceScaleFactor: 2 — retina-crisp text.
• printBackground: true — preserves colored backgrounds.
• margin: {top: 0,...} — no extra whitespace around slides.
• width/height must match --slide-w/--slide-h exactly.
• Wait for document.fonts.ready before capturing — Google Fonts must finish loading.

package.json setup

{
  "scripts": { "pdf": "node convert-to-pdf.js" },
  "dependencies": { "puppeteer-core": "latest" }
}

Use puppeteer-core (not puppeteer): the full puppeteer package downloads its own Chromium on install (~hundreds of MB) and uses that browser; puppeteer-core is the automation library only and expects you to pass executablePath to an existing Chrome/Chromium (this script finds system Chrome). Smaller install, one browser to update, fewer CI/cache issues.

Phase 6 — Verification

Screenshot each HTML slide and each PDF page, then compare visually.

Quick verification script

// Screenshot slide N from the HTML
const box = await page.evaluate((idx) => {
  const el = document.querySelectorAll(".slide")[idx];
  const r = el.getBoundingClientRect();
  return { x: r.x, y: r.y, width: r.width, height: r.height };
}, slideIndex);

await page.screenshot({ path: `html_slide_${n}.png`, clip: box });

What to check

• Page count matches slide count.
• No content clipped by footer on any slide.
• Fonts loaded (not fallback serif/sans-serif).
• All images render (no broken-image icons or alt text).
• No dark hue/shadow blocks around cards or images.
• Background colors and gradients preserved.
• Page edges are razor-sharp (no extra margin or whitespace).

Common pitfalls

Checklist

• Slide dimensions match --slide-w and --slide-h CSS variables
• Every .slide-content has bottom:52px to clear the footer
• @media print block strips all box-shadow values
• page-break-after: always set on .slide for print
• Converter uses puppeteer-core with system Chrome path
• printBackground: true and zero margins in PDF options
• deviceScaleFactor: 2 for retina quality
• Font loading awaited before PDF capture
• Image paths URL-encoded (no raw spaces)
• Visual verification done on cover, dense slides, and screenshot slides