For people already used to org-mode + curl (or any CLI tool), I highly recommend using those together when exploring an API (granted, this involves throwing simplicity out the window)

  :PROPERTIES:
  :header-args: shell :var PASSWORD=not :var USERNAME=narb :var AUTH_URL="nova.example.com:5000/v2.0/tokens"
  :END:
  ...
  #+NAME: get-openstack-token
  #+BEGIN_SRC shell :cache yes :results verbatim
  export DATA='{
      "auth" : {
          "tenantName": "'"$TENANT_NAME"'",
          "passwordCredentials": {
              "username": "'"$USERNAME"'",
              "password": "'"$PASSWORD"'"
          }
      }
  }'
  curl -s $AUTH_URL \
      -H "Content-Type: application/json" \
      --data "$DATA" | jq -r ".access.token.id"
  #+END_SRC

  #+RESULTS[be67...]: get-openstack-token
  : deadbeefdeadbeefdeadbeef

  ...
  #+BEGIN_SRC shell :var AUTH_TOKEN=get-openstack-token
  curl -H "X-AUTH-TOKEN: $AUTH_TOKEN"
  ...

Not to discount tools such as postman / insomnia or other built-in editor tools like restclient-mode, they are all fantastic tools

Some quick reasons why I found this works best for me:

* Org-mode is already great as a dedicated note-taking tool

* This is immediately compatible with existing CLI examples or output from tools that offer 'export as cURL' -- this is especially useful if I want to reverse engineer an api request by browser is making

* Easy to export to an html/wiki document for immediate consumption by others with working CLI examples

* Composable with other CLI tools like jq and xmllint

* Plain-text files are easy to version control