I use BlueEyes here (I'm a big fan of BlueEyes). It's JSON library is based on Lift-JSON. Lift-JSON is fairly commonly used, so perhaps you've run across it already. It's actually fairly simple to use and is one of the few parts of Lift that has decent documentation:

https://github.com/lift/lift/tree/master/framework/lift-base...

Here's a summary:

The JSON AST is pretty straightforward. The JSON DSL works like this:

- A tuple ("key", value) constructs a key/value pair

- ~ joins them together to construct an object

E.g. ("foo", 1) ~ ("bar", "Hi") = {"foo": 1, "bar": "Hi"}

I never actually invoke the parser myself. BlueEyes does it if you change the ByteChunk is my example above to JValue.

Drop me an email (noel at untyped dot com) or tweet (@noelwelsh) or IM or Skype or whatever if you want to discuss more.

Alternatively, I've never used Coda Hale's Jerkson package, but I expect it is of high quality.