markdown-to-pdf

name	markdown-to-pdf
description	Use when exporting Markdown reports to PDF, especially when Mermaid diagrams render blank, oversized, lose color, tables overflow, a TOC or PDF outline is needed, or Chrome/Puppeteer adds unwanted page headers and footers.
argument-hint	[source.md] [output.pdf]
allowed-tools	Bash, Read, Write
model	claude-sonnet-4-6
license	Apache-2.0
metadata	{"author":"Sagar Sarkale","version":"1.0"}

Rendering Markdown PDF

Overview

Markdown to PDF with Mermaid is a visual export task, not a plain file conversion. Render diagrams, style them consistently, disable print headers, then verify the PDF pages visually.

When to Use

Use for:

• Markdown docs containing Mermaid diagrams
• PDF export where diagrams are blank, oversized, clipped, or colorless
• Chrome/Puppeteer exports with unwanted page headers/footers
• report/document exports where tables and diagrams must stay readable

Do not use for plain text Markdown without diagrams.

Core Workflow

1. Preflight

   • Count Mermaid blocks in source.
   • Create a temp build dir.
   • Never edit source Markdown.

2. Render Mermaid explicitly

   • Use mmdc or browser-side Mermaid.
   • Apply a theme config with colors, font, and white background.
   • Set Mermaid flowchart.htmlLabels=false for PDF exports. Chrome PDF can drop foreignObject labels, causing blank boxes.
   • Wrap each diagram in a controlled container.

3. Build HTML with print CSS

   • Set @page { margin: 0.55in; }.
   • Add a designed Table of Contents near the top for long documents. Generate it from headings in the temporary build, not by editing source Markdown. Links are preferred when supported.
   • Add PDF outline/bookmarks for Preview/sidebar navigation. This is separate from visible TOC content. Use document headings, usually title + major sections.
   • Do not leave the TOC as raw Markdown bullets with awkward hyperlink styling. Render it as a polished nav.toc, ordered hierarchy, or table-style list with clean spacing, muted section numbers, and normal link text.
   • Use readable body width and table styles.
   • Start major sections on new pages. For docs where ## are major sections, use h2 { break-before: page; } and exempt the first major section if needed. For true H1 sectioned docs, use h1 { break-before: page; } except the document title.
   • Set diagram CSS:
     • max-width: 100%
     • height: auto
     • max-height: 7in
     • page-break-inside: avoid

4. Print with headers disabled

   • Prefer Puppeteer/CDP Page.printToPDF with displayHeaderFooter: false.
   • Do not rely on Chrome CLI defaults.

5. Verify

   • Confirm output PDF exists and has pages.
   • Rasterize PDF pages with pdftoppm or inspect screenshots.
   • Check diagram pages for blank placeholders.
   • Verify rendered diagram count equals Mermaid block count.
   • Check no page URL/date/title headers or footers exist.

Quick Reference

Symptom	Fix
Blank Mermaid block	Render with mmdc; verify SVG non-empty before PDF
Blank Mermaid node labels	Set flowchart.htmlLabels=false; avoid SVG foreignObject labels
Diagram too large	Strip fixed SVG dimensions; CSS max width/height
Mermaid colors gone	Use explicit Mermaid theme config and white background
Headers/footers on every page	Use Puppeteer displayHeaderFooter:false
Tables overflow	CSS table-layout:auto, smaller font, horizontal-safe widths
Major sections run together	Add print CSS page breaks on h1/h2, excluding title
Long doc hard to navigate	Generate TOC from headings in temp HTML/PDF build
Preview sidebar index missing	Add PDF outline/bookmarks from headings; visible TOC alone is not enough
TOC looks like bullet links	Replace raw ul bullets with styled nav.toc, ordered hierarchy, or table-style rows
Source got changed	Stop. Restore source. Use temp build artifacts only

Navigation Requirements

Long PDFs need two navigation layers:

1. Visible TOC page/section inside PDF content.
2. PDF outline/bookmarks visible in Preview, Acrobat, browser sidebars, and other PDF readers.

Do not treat these as interchangeable. A clickable TOC page does not create Preview sidebar entries.

Visible TOC Style

For long PDFs, add a Table of Contents that looks intentional:

• Place TOC after the document title and any short prefatory note.
• Use heading text from source, but generate TOC in temp HTML/build artifacts only.
• Prefer a nav.toc block with title Table of Contents.
• Avoid default Markdown bullets. If using ul/ol, set list-style: none and create visual hierarchy with spacing/indentation.
• Avoid ugly blue underlined links. Use document text color, subtle hover/print styling, and no text decoration in print.
• Make each entry feel like one clean row: section number or level marker, title, optional page/anchor affordance.
• Include top-level sections by default. Include second-level sections only when document is short enough or nested indentation stays readable.
• Keep TOC compact: no more than two pages unless user asks for exhaustive navigation.

Example CSS pattern:

.toc {
  margin: 1.5rem 0 2rem;
  padding: 1rem 1.25rem;
  border: 1px solid #d8e0ee;
  border-radius: 12px;
  background: #f8fbff;
}
.toc h2 {
  margin: 0 0 0.75rem;
  break-before: auto;
}
.toc ol {
  list-style: none;
  margin: 0;
  padding: 0;
}
.toc li {
  margin: 0.35rem 0;
  line-height: 1.35;
}
.toc a {
  color: #162033;
  text-decoration: none;
}
.toc .toc-level-2 {
  margin-left: 1rem;
  color: #526173;
  font-size: 0.95em;
}

PDF Outline / Bookmark Requirements

Add document outline entries from headings:

• Include document title.
• Include major sections by default: usually h1 title and h2 sections, or h1 sections when a doc uses multiple top-level headings.
• Include h3 only when nesting stays readable.
• Keep labels clean: remove Markdown numbering artifacts if duplicated, trim whitespace, decode HTML entities.
• Bookmarks must jump to correct section pages, not just page 1.
• Prefer PDF generators that emit outlines natively. If using Puppeteer/Chrome and it does not emit outlines, postprocess the PDF with a PDF library that can add outlines/bookmarks.
• If outline creation is unsupported by the available toolchain, say so explicitly in the output report instead of implying the TOC covers it.

Validation options:

# If mutool is available
mutool show output.pdf outline

# If qpdf is available, inspect JSON for outline data
qpdf --json output.pdf | rg -i 'outline|bookmark|title'

Validation Commands

# Count diagrams in Markdown
rg -n '^```mermaid' input.md

# Check PDF page count
pdfinfo output.pdf | rg '^Pages:'

# Rasterize for visual inspection
mkdir -p /tmp/pdf-pages
pdftoppm -png -r 120 output.pdf /tmp/pdf-pages/page
open /tmp/pdf-pages

Recommended Toolchain

Preferred:

Markdown -> temp Markdown/HTML -> Mermaid SVG -> styled HTML -> Puppeteer PDF

Use:

• mmdc for Mermaid SVGs
• pandoc or Python Markdown for HTML
• Puppeteer/Chrome CDP for PDF
• pdfinfo and pdftoppm for verification

Common Mistakes

• Assuming successful PDF creation means diagrams rendered.
• Trusting Chrome CLI --print-to-pdf without checking headers/footers.
• Letting SVG intrinsic dimensions control PDF layout.
• Leaving Mermaid source blocks after failed rendering.
• Allowing Mermaid foreignObject HTML labels in PDF exports. They may show in browser but print as blank boxes.
• Verifying only the first page.
• Forgetting that many Markdown docs use ## as major sections after one document # title.
• Leaving auto-generated TOC as raw bullet points with blue underlined links.
• Assuming a visible TOC creates PDF bookmarks. Preview sidebar index requires PDF outline entries.
• Editing the source Markdown to make PDF conversion easier.

Required Output Report

After conversion, report:

• Output path
• File size
• Page count
• Mermaid block count vs rendered SVG count
• Whether headers/footers are absent
• Whether any diagram pages were visually checked
• Whether major sections start on new pages
• Whether TOC was generated
• Whether PDF outline/bookmarks were generated and how verified
• Toolchain used