Warn and repair invalid UTF-8 content

The API will reject API calls with invalid UTF-8 content with a 406 "Not Accepted" error, but this could be avoided by scrubbing the input before sending it.

There seems to be a way to do this using [Buffer](http://stackoverflow.com/questions/17006799/how-do-i-capture-utf-8-decode-errors-in-node-js) to force-encode things.