host-pdf

host-pdf: Export HTML Slides to PDF

When the user runs /host-pdf, follow these steps:

Step 1: Find the HTML slide deck

Look for the HTML file to export. Check in this order:

1. If the user specifies a file path (e.g., /host-pdf presentation.html), use that file
2. If there's a recently generated/modified .html file in the current project, use that
3. Ask the user which HTML file to export

The HTML file must contain slide elements with the class .slide — this is the standard format produced by host-html slide decks.

Step 2: Ensure Puppeteer is available

Check that Puppeteer and its Chrome browser are installed:

npm ls puppeteer 2>/dev/null || npm install puppeteer
npx puppeteer browsers install chrome 2>/dev/null

These commands are safe to run even if already installed.

Step 3: Run the export script

The export script is bundled with this skill. Run it:

node "${CLAUDE_PLUGIN_ROOT}/skills/hosthtml-pdf/export-pdf.mjs" "<INPUT_HTML>" "<OUTPUT_PDF>"

• ${CLAUDE_PLUGIN_ROOT} is the root directory where the plugin is installed (e.g., ~/.claude/plugins/cache/host-html)
• <INPUT_HTML> is the absolute path to the HTML file from Step 1
• <OUTPUT_PDF> (optional) defaults to the same name/location as the input but with a .pdf extension

Optional flags:

• --slides N — override the auto-detected slide count (useful if slides don't use the .slide class)

Example:

node "${CLAUDE_PLUGIN_ROOT}/skills/hosthtml-pdf/export-pdf.mjs" ./deck.html ./deck.pdf

Step 4: Present the result

Once the script finishes, show the user:

1. PDF path: The absolute path to the generated PDF
2. Slide count: How many slides were captured
3. File size: The size of the generated PDF

Format the output clearly:

PDF exported!

File: /path/to/deck.pdf
Slides: 8
Size: 2.4 MB

Error Handling

• "No .slide elements found": The HTML doesn't use .slide class divs. Ask the user how many slides to capture and retry with --slides N
• Puppeteer launch failure: Suggest running npx puppeteer browsers install chrome to install the browser
• Timeout or blank captures: The HTML may need more time to render — the script already waits 2s for initial load and 900ms per slide, which handles most cases

Tips

• The script captures at 2× device scale (retina quality) at 1280×720 — the PDF will be crisp
• Each slide is screenshot-based, so all CSS animations, custom fonts, and backgrounds are preserved exactly as rendered
• If the user wants to share the PDF online, suggest combining with /host-html to also publish the live version