Alternatives · Puppeteer

From wkhtmltopdf to Puppeteer: what changes

Puppeteer, the Node.js library that drives Chrome (Apache-2.0), prints with page.pdf(); wkhtmltopdf’s status page suggests it for pages with dynamic JavaScript. pdfkiln prints with the same Chromium, but answers to wkhtmltopdf’s command line.

What changes when you leave wkhtmltopdf for Puppeteer

• Code instead of a command line. A Node.js program launches the browser, opens the page and calls page.pdf(). The libraries that call wkhtmltopdf, knp-snappy, pdfkit, wicked_pdf and the others, do not call Puppeteer.
• Options. wkhtmltopdf’s options become Puppeteer’s PDF options (format, margin, scale, printBackground, pageRanges…) or code; the cover page and the table of contents have no equivalent.
• Headers and footers. Puppeteer’s are HTML templates filled through five classes (date, title, url, pageNumber, totalPages). wkhtmltopdf’s --header-html and --footer-html are full pages, with page, topage, section and others in their address, and their JavaScript: templates built on them, Odoo’s and Frappe’s among others, are to be rewritten.
• Layout. Chrome lays out the page as a browser does, without wkhtmltopdf’s quirks (the 0.8 shrinking, @page ignored, flex and grid ignored…): templates tuned for wkhtmltopdf can change page breaks and page count.
• The browser. Installing Puppeteer downloads Chrome for Testing, about 282 MB on Linux, and chrome-headless-shell.
• Internal addresses. The page loads what it asks for; keeping it away from internal addresses, the problem of CVE-2022-35583, is up to your code, through request interception.

When Puppeteer is the right choice

New Node.js code, or more than PDFs: screenshots, tests, scraping, a page that needs a login or a click before it prints.

When pdfkiln is

Code or libraries that call wkhtmltopdf today, templates whose page count must not change, another language than JavaScript, no browser automation to write: pdfkiln takes wkhtmltopdf’s place, Chromium underneath, with the emulations that keep its pages. Called as pdfkiln, it gives Chrome’s current CSS without them.