Jupyter Notebook to PDF: 4 Methods Compared

There are multiple ways to export a Jupyter Notebook to PDF, and each has real tradeoffs. This article compares the four main methods in depth so you can choose the right one for your situation.

Overview

Method	LaTeX Required	Installation	Quality	Privacy	Automation
nbconvert (LaTeX)	Yes	Yes	Excellent	Local	Yes
nbconvert (WebPDF)	No	Yes (Playwright)	Very Good	Local	Yes
Browser Print	No	No	Very Good	100% Local	No
Online converter	No	No	Very Good	Varies	No

Method 1: nbconvert with LaTeX (Classic)

This is the official, standard method that ships with Jupyter.

pip install nbconvert
jupyter nbconvert --to pdf my_notebook.ipynb

Under the hood, this converts your notebook to a .tex file and then compiles it with xelatex. The resulting PDF has excellent typographic quality — proper page layout, fonts, and consistent formatting.

Installation Requirements

You need:

Python with jupyter and nbconvert installed
A full LaTeX distribution:
- macOS: MacTeX (4 GB download from tug.org/mactex)
- Ubuntu/Debian: sudo apt-get install texlive-xetex texlive-fonts-recommended texlive-latex-extra
- Windows: MiKTeX or TeXLive

Total disk space: 2–8 GB depending on the distribution.

Pros

Highest typographic quality
Supports complex math through native LaTeX rendering
Produces a proper "document" feel with consistent fonts and layout
Works from the command line — easy to script and automate
Part of the official Jupyter ecosystem

Cons

LaTeX installation is large, complex, and frequently causes errors
Error messages are cryptic (! LaTeX Error: File tcolorbox.sty not found)
Can fail with Unicode characters, wide tables, or certain output types
Slow — LaTeX compilation can take 30+ seconds for large notebooks
Requires a working Python environment

Best For

Academic papers and formal reports that need LaTeX-quality typography
Automated PDF generation in CI/CD pipelines (once LaTeX is set up)
Notebooks with complex mathematical notation

Method 2: nbconvert with WebPDF (No LaTeX)

Introduced in nbconvert 6.x, the webpdf exporter uses a headless Chromium browser instead of LaTeX.

pip install "nbconvert[webpdf]"
playwright install chromium
jupyter nbconvert --to webpdf my_notebook.ipynb

How It Works

The webpdf exporter:

Converts the notebook to HTML (same as --to html)
Opens the HTML in a headless Chromium browser
Uses Chromium's built-in print-to-PDF functionality
Returns the resulting PDF

Installation Requirements

Python with nbconvert[webpdf] extra
Playwright (Python package)
Chromium browser (~150 MB download, handled by Playwright automatically)

Pros

No LaTeX required
Better handling of HTML output (DataFrames, plotly charts, rich output)
Still scriptable from the command line
Produces results that match what you see in the browser

Cons

Still requires Python and a Chromium download
Marked as experimental in some versions
Slightly slower than LaTeX for text-heavy notebooks (Chromium startup overhead)
Less typographic polish than the LaTeX pipeline for pure-text content

Best For

Command-line automation without LaTeX
Notebooks with complex HTML outputs (interactive widgets frozen, Plotly charts)
Teams that want to avoid LaTeX but still want CLI automation

Method 3: Browser Print (Manual)

The simplest approach that requires no installation at all.

Option A: Export to HTML, then Print

Export to HTML with nbconvert:

jupyter nbconvert --to html my_notebook.ipynb

Open my_notebook.html in Chrome or Firefox
Press Ctrl+P (Windows/Linux) or Cmd+P (Mac)
Set destination to "Save as PDF"
Click Save

Option B: Use an online converter

Tools like notebookconvert.com do the HTML rendering step for you in the browser — just upload your .ipynb file directly.

How It Works (for online tools)

A client-side JavaScript converter:

Reads the .ipynb JSON file
Parses cells (code, markdown, raw)
Renders code cells with highlight.js
Renders markdown with marked.js
Renders equations with MathJax (if detected)
Embeds all images as base64 data URIs
Opens a print dialog
You save as PDF using the browser's PDF print driver

Pros

Zero installation
Works on any device with a modern browser (including Chromebooks, tablets)
100% private — files never leave your computer (for client-side tools)
Fast — no compilation step
Handles Unicode, emojis, and any text content
Works even if your Python environment is broken

Cons

Cannot be scripted or automated
Print dialog is a manual step
Less typographic control than native LaTeX (fonts, exact page layout)
PDF quality depends on the browser's print engine

Browser-Specific Tips

Chrome/Edge:

Enable "Background graphics" in "More settings" to keep syntax highlighting colors
Set margins to "Minimum" for a cleaner layout

Firefox:

Under "More Options", check "Print Background Colors and Images"
Firefox's PDF output is slightly different in page sizing

Safari:

Use File → Export as PDF rather than the print dialog for better results

Best For

One-off conversions with no setup
Users who don't have Python installed
Sharing with colleagues quickly
Notebooks with rich HTML outputs

Method 4: VS Code Export

If you use VS Code with the Jupyter extension (ms-toolsai.jupyter), you can export to PDF directly from the editor.

Open your .ipynb file in VS Code
Click the ... (More Actions) button in the notebook toolbar
Select Export → PDF

VS Code uses a headless browser internally (similar to Method 2).

Pros

Integrated into your editor workflow
No terminal needed
No separate LaTeX installation

Cons

Requires VS Code with the Jupyter extension
Less configurable than the CLI approaches
Quality is similar to the browser-print approach

Best For

VS Code users who want a quick export without switching to the terminal

Which Method Should You Use?

For a quick one-off export

Use notebookconvert.com or the browser print method. Zero setup, works anywhere, and the quality is excellent for most use cases.

For automated pipelines (CI/CD)

Use nbconvert --to webpdf if you want to avoid LaTeX, or nbconvert --to pdf if you need LaTeX-quality output and are willing to set up a LaTeX installation in your CI environment.

Example GitHub Actions step for webpdf:

- name: Install nbconvert with webpdf
  run: pip install "nbconvert[webpdf]" && playwright install chromium

- name: Convert notebook to PDF
  run: jupyter nbconvert --to webpdf report.ipynb

For academic submissions with strict formatting requirements

Use the LaTeX method (nbconvert --to pdf). It produces the most professionally typeset output and is the most widely accepted in academic contexts.

For sharing reports with non-technical stakeholders

Both the HTML and PDF browser-based approaches work well. HTML is often better for interactive content; PDF is better for printing or archiving.

Quality Comparison

The quality difference between methods is subtle for most notebooks. Here's what to expect:

Code formatting: All four methods produce syntax-highlighted code. The LaTeX method uses the listings or minted package; the browser methods use highlight.js. The visual result is similar.

Math equations: LaTeX native rendering is slightly higher quality than MathJax. For most equations, the difference is not visible. For complex multi-line equations, LaTeX handles alignment better.

Charts and figures: Browser methods embed PNG/SVG as-is, so the quality is exactly what you see in the notebook. LaTeX converts everything to a format it can embed, which may rasterize SVGs.

Text layout: LaTeX's typographic algorithm (hyphenation, line breaking, kerning) is superior for long paragraphs of text. Browser PDF printing is excellent for code-heavy notebooks.

Conclusion

There is no single "best" method — it depends on your workflow. For most data scientists who need to share a notebook quickly, the browser-based approach (online tool or browser print) is the fastest path. For teams with serious automation needs or academic formatting requirements, the LaTeX or webpdf pipelines are worth the setup investment.

The good news is that with modern browsers and tools like notebookconvert.com, you don't need to fight with LaTeX for the common case.