HTML to PDF API: hosted services, the headless Chrome route, and when a template API is better

What an HTML to PDF API is, and the four hosted ones compared here

An HTML to PDF API accepts HTML, or a URL, in an HTTP request and returns a rendered PDF, so you do not run a browser or a rendering engine on your own servers. The four hosted services below all work that way. They differ in the engine behind them, in how they count usage and in price.

There is also a free route: run headless Chrome yourself. It costs nothing in license fees and is a fair choice for many teams, so it gets its own section.

DocRaptor, PDFShift, Api2Pdf and Doppio: calls and prices

Endpoints and parameters come from each vendor's documentation, and prices from each pricing page, all read on 3 October 2026. Prices are monthly billing in US dollars.

Counting rules matter as much as the headline price. DocRaptor charges the same for any document size. PDFShift counts one credit per 5 MB of output. Api2Pdf bills bandwidth and compute time. PDFShift's documentation also says requests without a valid key are treated as a free account: watermarked, capped at 2 MB and 10 conversions a day.

The open source route: Puppeteer, Gotenberg and why not wkhtmltopdf

If you can run Chrome, you do not need to pay for conversion.

• Puppeteer (Node.js). Install with npm i puppeteer, which also downloads a Chrome for Testing build of roughly 170 to 282 MB depending on the platform. Load your markup with await page.setContent(html) and write the file with await page.pdf({ path: 'out.pdf' }).
• Gotenberg (any language). Start the container with docker run --rm -p "3000:3000" gotenberg/gotenberg:8, then POST a file named index.html to /forms/chromium/convert/html. You get an HTTP API that you host.
• wkhtmltopdf. Still widely installed, but its own status page says the WebKit it uses has not been updated since 2012 and warns against feeding it untrusted HTML. That page points report builders to WeasyPrint and pages with dynamic JavaScript to Puppeteer.

The step-by-step guide to generating a PDF from HTML code walks through the Puppeteer calls in order.

When a hosted API beats running Chrome yourself

Self-hosting is free until it is not. Browsers are heavy processes, each render holds memory, and someone has to patch the browser and size the queue. A hosted API is worth paying for when PDF volume is bursty, when your platform cannot run a browser at all, or when you need print features a browser handles poorly. DocRaptor's own case for its Prince engine is exactly that: complex layouts, tables of contents, and headers or page sizes that vary within one document.

If you already run containers and produce a few hundred simple documents a month, Gotenberg or Puppeteer is enough and you should not buy anything. Whatever you choose, test long tables early, because a table that splits across pages needs explicit CSS; see the page breaks guide.

When a template API is the better fit than HTML to PDF

HTML to PDF assumes a developer owns the layout. Every wording change, logo swap or extra column is a code change and a deploy. A template API moves the layout into an editor and leaves your code sending JSON only.

CraftMyPDF is the example this site covers most closely. Its editor is drag and drop and needs no HTML or CSS, and its Free plan allows 50 PDFs a month with 3 templates. APITemplate.io sits in between, with both a visual editor and an HTML editor. Choose that family when a non-developer must be able to change the document. Stay with HTML to PDF when your HTML already exists and changes rarely. The PDF generation API comparison lists both families side by side, and CraftMyPDF vs DocRaptor compares one of each directly.