[benibela]

You could try my Xidel, based on XPath and JSONiq ( http://www.videlibri.de/xidel.html ).

It is more verbose for simple queries, but with named variables and functions, larger queries do not become complicated

[inshadows]

Xidel is my immediate pick for querying HTML. I use it in several scripts. Thanks!

I still prefer jq for JSON processing though. I'm hooked to pipelines. EDIT: The great power of jq is in transforming documents, not querying.