Hi Musaffir,
I blogged about how we chose an API tool on my blog, here: http://testernaut.blogspot.com/2018/05/choosing-api-tool.html - this speaks to how we chose, languages etc.
Shortly after this, we found that frisby.js had some flaws and we switched to IcedFrisby.
There was a feature we needed for a particularly rough API that required cookie authentication (ew), which required some changes to IcedFrisby to support. We could have implemented the bits we needed via some hacks our side (probably), but contributing to the repo seemed like the best way, and one of the devs was encouraging me to get involved in Hacktoberfest. So I did, and I improved some docs and other things whilst I was there (and earned my Hacktoberfest t-shirt!)
The company now has a very decent collection of API tests, and Iโm a maintainer on the repo.
On your question about CI: we used a CI system to monitor our live environment using a subset of the tests (because regularly testing login and other business logic operations is way more effective and informative than pinging). We wanted to get to what I think youโre talking about though - builds being automatically deployed and tested, but we got scuppered on the โdeployโ part - it was non-trivial, and wasnโt worth the investment whilst the company was on a path towards .NET Core and Docker.
Hope this helps!
Dan