When a service needs an account per user, the agent can have one of its own. Those accounts are created by a live signup in a real browser on your machine.
Because that is what the service expects. A real signup form, filled in a real browser, with a real verification code arriving in a mailbox the agent can read. No API shortcut exists for most of these services.
Usually an email address, sometimes a phone number. If the signup needs one and the agent has none, the signup waits for you rather than inventing something.
Some services actively resist automated signup. A signup can fail for reasons that have nothing to do with your setup: network reputation, volume from an address range, a service treating a new mailbox as disposable, or a challenge that requires a human. Those are honest failures, not misconfiguration. When a signup fails explains what is worth retrying.
The account exists on that service, owned by you through the agent. You can read its login with a reveal link, and the account keeps existing even if you later remove the agent.