Onyx feels more like a library then a framework, so it's not as 'heavy' as other frameworks. Also, since it relies on pure data and pure functions, it was actually trivial to get up and running with it when I was evaluating it. There was about 20 lines of boilerplate I needed to copy, and then I just built a workflow with the functions I had already been using and it all just kinda worked.

I was able to develop in the REPL, which was a huge win, and deploying is as simple as uberjar / copy to host.