aws-samples Read a PDF file
Converts a PDF to Markdown and prints to stdout:
uv run --with pymupdf4llm ~/.kiro/skills/pdf-to-markdown/scripts/pdf_to_markdown.py "<path_to_pdf>"
Save as .md file:
uv run --with pymupdf4llm ~/.kiro/skills/pdf-to-markdown/scripts/pdf_to_markdown.py "<path_to_pdf>" "<output.md>"
Write a PDF file
Preferred: Directly from markdown content (no intermediate file)
Pipe markdown content directly using --stdin:
echo '<markdown_content>' | uv run --with pymupdf ~/.kiro/skills/pdf-to-markdown/scripts/markdown_to_pdf.py --stdin "<output.pdf>"
For multi-line content, use a heredoc:
uv run --with pymupdf ~/.kiro/skills/pdf-to-markdown/scripts/markdown_to_pdf.py --stdin "<output.pdf>" <<'EOF'
# Title
Some paragraph with content.
- Bullet point 1
- Bullet point 2
1. Numbered item
2. Another item
EOF
From an existing Markdown file
uv run --with pymupdf ~/.kiro/skills/pdf-to-markdown/scripts/markdown_to_pdf.py "<input.md>" "<output.pdf>"
If you had to create the .md file as an intermediate step, delete it after generating the PDF.
Notes
- Read uses pymupdf4llm for high-quality text extraction preserving structure, tables, and image references.
- Write uses pymupdf (fitz) to generate styled A4 PDFs with headings, bullets, numbered lists, blockquotes, and horizontal rules.
- Dependencies are auto-installed via uv run.
- IMPORTANT: Do NOT create intermediate markdown files if you can pipe content directly via
--stdin. If you must create one, always delete it after.