wonderful. That's a little non-intuitive that by setting the accept header I end up cleaning up the data, but in hindsight it makes sense. Maybe a little callout in the manual?
Earlier during our dev phase, we made cleanup the default but realized that devs would want raw output in all cases unless they specifically asked for this behavior. We encoded that specific knowledge in the accept header. I agree we could have done it differently but we wanted to stay away from adding another method to the "y" global object.
Totally agree about the documentation and this is being worked on as we speak. -- Nagesh