The documentation about WolframLanguageForJupyter, by Wolfram Research, available at https://github.com/WolframResearch/WolframLanguageForJupyter, gives two methods for installing this interface: Method 1: using wolframscript; and Method 2: Using Wolfram Language.
For Method 2, one is supposed to download and install a paclet of the form WolframLanguageForJupyter-x.y.z.paclet. But the link https://github.com/WolframResearch/WolframLanguageForJupyter/releases there has target a page with no releases!
I did download and unarchive WolframLanguageForJupyter-master.zip from the Github project page, but although there is a PacletInfo.m file there, still I find no actual paclet.
Where is the paclet?
I could not find the paclet as well. Here is an alternative way for the installation with the m–file that is attached in the download from Github.
I'm using macOS.
In any case, I think pip is to be avoided, given that the Python I use is either from within Conda/Anaconda or else the MacPorts version.
Did you read the "Instructions for macOS" at the end of the suggested stackexchange website?
Yes, I did read "Instructions for macOS". (Apparently the link in your initial answer here points to the answer for Windows.)
The method does work just fine.
Still puzzling why no paclet is on the github repository yet the docs on the project's page (and its readme file) describe the use of such!
Using paclets is a much handier means of installing things because one can subsequently use PacletUpdate (or even PacletUninstall) from PacletManager.
I filed a bug report at the GitHub project page. The paclet has now been added there, and the paclet method of installing works.
It works!! Not perfectly, but the fact that it works at all -- indeed, pretty well at first glance -- is amazing. I believe this is a major development for the Wolfram Language. Yes, the Jupyter frontend is way less sophisticated than the Mathematica front end, but the fact that millions of people are using it and developing for it is a major plus. To see a small sample of what it can do -- and I literally created this in about half an hour -- check out this link: HTML export of Jupyter Notebook sitting in Wolfram Cloud.