Hi Scott - hit a bone of contention there.
What we can offer today is the Structure component detailed here - http://www.scryber.co.uk/HowTo/tabid/...
This will accept a string of scryber components that can be data-bound in the example. So you can take your xhtml, transform it as required and set the value.
As for input of HTML - I actually don't think we will make this for 0.9 as a final version. There are too many queries such as support for CSS inline styles, CSS referenced styles, images within CSS, invalid HTML, floating content, absolute and relatively positioned elements... the list is endless.
We may get to a point where well formed XHTML snippets are supported and maybe CSS referenced styles but not at the moment. Could not support the vague-ness to end producers.
I hope this at least explains the situation. For the moment, if you can convert the HTML to PDFX then bring it in with the Structure.