Your first collection

POST /api/collections defines a collection’s fields and access rules; Cratebase creates and migrates the real SQL table for you (a cb_<name> table with one physical column per field — see Collections):

curl -X POST localhost:8090/api/collections \
  -H "authorization: Bearer $ADMIN_TOKEN" -H 'content-type: application/json' \
  -d '{"name":"posts","type":"base",
       "schema":[{"id":"f1","name":"title","type":"text","required":true},
                 {"id":"f2","name":"published","type":"bool"}],
       "listRule":"","viewRule":"","createRule":"","updateRule":"","deleteRule":null}'

curl -X POST localhost:8090/api/collections/posts/records \
  -H 'content-type: application/json' -d '{"title":"Hello","published":true}'

curl "localhost:8090/api/collections/posts/records?filter=published%20%3D%20true&sort=-created"

The same thing works from the dashboard: Collections → New collection walks through the same fields and rules, and the per-collection API tab shows the request shapes for whatever you just defined.