I did something similar (http://markolson.github.io/storyboard/) earlier this year using ffmpeg and ImageMagick to generate either GIFs around lines of dialog (like this project), or PDFs where each page is a frame of text or a new scene. Optimizing GIFs is by far the least enjoyable part.

Damn you I was going to come post your storyboard. :(

Yahoo and others either rate limit or do temporary IP bans if you access too many pages too quickly. Distributing the tools through a VM distributes the workload in a pretty easy-to-setup way.

Are you looking for docs on Chef-Solo, or setting up multiple VMs to use as a Chef Server and additional nodes?