An HtmlRenderer.PdfSharp alternative

HtmlRenderer.PdfSharp is a free, managed HTML to PDF converter that runs on top of PDFsharp. Two things make people look for a replacement:

The licence of the engine underneath. HtmlRenderer.PdfSharp 1.6.1 itself is BSD-licensed and still installs PDFsharp 6.2.4, the last MIT version. PDFsharp 6.5.0 (October 2026) and the 7.0 previews are AGPL-3.0 or a paid commercial licence. Updating PDFsharp, by hand or with a tool such as Dependabot, adding another package that needs a newer PDFsharp, or a future HtmlRenderer release built on it, brings in the AGPL, which requires you to publish your own application's source code unless you buy a licence. Staying on MIT means pinning PDFsharp at 6.2.4 and doing without its future fixes.
The CSS. HtmlRenderer supports HTML 4 and CSS 2, so flexbox, grid, CSS variables, calc() and @page rules are ignored.

PdfPecker is also fully managed and runs in your own process, with no dependencies at all, under the MIT licence. It supports modern CSS and has PDF/A, PDF/UA, signatures, forms and merging built in.

dotnet add package PdfPecker

Then remove the HtmlRenderer.PdfSharp package, and PDFsharp too if nothing else in your project uses it.

Replace the HtmlRenderer call

// Before
using PdfSharp;
using TheArtOfDev.HtmlRenderer.PdfSharp;

var document = PdfGenerator.GeneratePdf(html, PageSize.A4);
document.Save("invoice.pdf");

// After
using PdfPecker;

File.WriteAllBytes("invoice.pdf", HtmlToPdf.Convert(html));

A4 is the default. With margins and a page footer:

// Before
var config = new PdfGenerateConfig { PageSize = PageSize.A4, MarginTop = 40, MarginBottom = 40 };
var document = PdfGenerator.GeneratePdf(html, config);

// After
byte[] pdf = HtmlToPdf.Convert(html, new PdfOptions
{
    Margins = PageMargins.FromMillimeters(14, 15, 14, 15),
    FooterText = "Page {page} of {pages}",
});

HtmlRenderer has no page numbers of its own; PdfPecker has FooterText, @page margin boxes and running headers and footers written in HTML (see Paged media).

Options side by side

HtmlRenderer.PdfSharp	PdfPecker
`PageSize`, `PageOrientation` in `PdfGenerateConfig`	`PageSize = PageSize.A4.Landscape`, or CSS `@page { size: A4 landscape }`
`MarginTop`, `MarginBottom`, `MarginLeft`, `MarginRight` (points)	`Margins = new PageMargins(top, right, bottom, left)` in points, or `PageMargins.FromMillimeters`
`cssData` argument	`UserStyleSheet`, or `<style>` and `<link>` in the HTML
`imageLoad` and `stylesheetLoad` events	`ResourceResolver`, or files under `BaseDirectory`
`PdfDocument` for further editing	PDF bytes; merge, stamp, sign and protect them with PdfPecker's own APIs
Fonts from PDFsharp's `GlobalFontSettings.FontResolver`	`options.Fonts.AddFile(...)` or `@font-face` (see Fonts)

What changes in the output

Modern CSS now works. Layouts built with tables and floats to suit CSS 2 keep working, and you can move to flexbox and grid when you want.
Backgrounds and colours that HtmlRenderer dropped, such as coloured table headers, are drawn.
Long tables break across pages with their header row repeated, and rows are never cut in half.
Text is laid out by PdfPecker's own engine, so line breaks and page breaks can fall in slightly different places. Check your longest templates after switching.

Speed and memory

On the same 4-core Linux VM, a one-page invoice took 6.3 ms with HtmlRenderer.PdfSharp and 0.86 ms with PdfPecker, a 30-page report 639 ms and 45 ms, and peak memory was 236 MB and 90 MB. See Performance.

When HtmlRenderer is still a fine choice

For simple HTML that already looks right, there is no urgent reason to move today: version 1.6.1 runs on the MIT PDFsharp 6.2.4. Pin that version so an update can't change the licence. PDFsharp itself remains a solid API for drawing PDFs from code. Compare shows the two side by side with the other .NET converters.