Looks like a potential opportunity to use something like node-unfluff [0] to scrape page content. Pull request anyone?

[0]: https://github.com/ageitgey/node-unfluff