I used this for a major company's site-edit auditing system. (No, they didn't want HTML snapshots of each revision. It had to be a screenshot of the browser...)
It works really well. The only quirk is that it needs a fake X server (for font loading), but Xvfb works just fine for that.
I did not know about this project at the time I started with pdfcrowd. But anyway, I just took my existing pdf library and integrated it with WebKit which was not that hard as one could think.
First of all, I don't know how well wkhtmltopdf works, but there are many, many solutions to the HTML-to-PDF problem, and most of them suck. It's not surprising the creator decided to put together a library from scratch, it's the special sauce for his business.
Also, the "value add" comes from the fact that wkhtmltopdf is a library, and PDFcrowd is an API.
It works really well. The only quirk is that it needs a fake X server (for font loading), but Xvfb works just fine for that.