rustedf.cloud
Menu

HTML to PDF in C#: PdfSharp, QuestPDF, iText, Puppeteer or an API

Published 4 min read

The first decision in C# is not which library to install, but which kind of tool you want. They differ in what you write (code, HTML or a layout), where the work happens (inside your process, in a browser you ship, or on someone else’s server) and what you have to keep up to date. This guide sorts the options by kind. We make one of them, the API, so we say plainly where it is not the right answer.

Libraries change their features and licenses. Read the current documentation and license of anything you consider; what follows describes the approach of each, not the fine print.

Six approaches

Approach Examples You write Watch for
Draw from code PdfSharp, MigraDoc Drawing at positions you choose (PdfSharp) or a document model with flowing text and tables (MigraDoc), all in C#. No HTML conversion of its own: HTML has to be turned into this model by you or by an add-on.
A layout in C# QuestPDF A fluent C# description of the document, with a real layout engine. Not HTML. Its license depends on the size and kind of your company: read the current terms.
HTML through a library iText with its HTML add-on, packages built on HtmlRenderer HTML and CSS, converted in your process. iText is dual-licensed (open source AGPL or commercial). HtmlRenderer targets HTML 4.01 and CSS 2, so modern layout such as flexbox is not what it is for.
A browser you run PuppeteerSharp, Playwright HTML, CSS and JavaScript, rendered by Chromium. You download and keep up to date a browser, with its memory, start-up time and sandbox.
An old converter wkhtmltopdf and its wrappers (DinkToPdf, Rotativa) HTML and CSS. The upstream project is archived: no more fixes. Its own documentation warns against untrusted HTML.
An HTTP API rustedf.cloud HTML and CSS, or a saved template, sent with HttpClient. Your data leaves your server. It needs the network, and your plan sets the limits.

How to choose

Ask these questions in order. The first honest “yes” usually decides it.

Does the page need JavaScript to exist? A single-page app, a chart drawn in the browser, a page that builds itself on load. Only a browser renders that. PuppeteerSharp or Playwright are the answer, and the API is not.

Is the document built in C# and has no HTML? Invoices assembled from objects, labels, tickets. A layout library such as QuestPDF or a drawing library such as PdfSharp fits, and they need no network.

Can the data leave your server? If a regulation or a contract says it cannot, use something that runs in your process or in a browser you host. An API call sends the document’s data to a service.

Do you already have HTML? A Razor view, an email template, a page built for printing. Then the choice is between a browser and the API. The browser has the full set of web features and costs you its upkeep. The API has none of the upkeep and supports the parts of HTML and CSS that matter for paper: pages, tables that continue across pages, headers and footers.

Where does it run? In a container or a serverless function, a Chromium adds to the image and to the cold start. An HTTP call adds nothing to either.

What the API version looks like

using var response = await http.PostAsJsonAsync(
"v1/render", new { html, title = "Invoice 2026-0042" }, cancellationToken);
response.EnsureSuccessStatusCode();
byte[] pdf = await response.Content.ReadAsByteArrayAsync(cancellationToken);

There is no package to install and no native file to ship. The cost is the list of what it does not do: no JavaScript, no @font-face or @import, and images only as data: URLs or files you upload. Send the same body to POST /v1/validate first: it tells you what the service could not use, with no PDF and no credit.

Where to go next

Read Supported HTML and CSS before you decide: it lists exactly what the renderer supports.

← All posts