Email quality

When you pass an email to an evaluation, Rupt inspects the address and the domain behind it. None of these checks judges a person, they judge the address, but together they're the backbone of fake-account detection, because bulk registration almost always cuts corners on email.

The checks

• email_is_disposable: the domain is a throwaway provider built to self-destruct (Mailinator, 10minutemail, and thousands like them). A real customer rarely signs up with one; a bot farm almost always does.
• email_is_invalid: the domain can't actually receive mail (no valid MX records) or the address is malformed. Often a typo, sometimes a fabricated address.
• email_is_webmail: a free consumer provider like Gmail or Outlook. This is the weakest of the set, since most real people use webmail. It only carries weight stacked with other signals.
• email_is_accept_all: the domain accepts mail to any address, so you can't tell a real mailbox from an invented one. Common with catch-all business domains, which makes it ambiguous rather than damning.
• is_email_verified: whether the user has confirmed they control the inbox. An unverified address is just a claim.

Using it

Each check is a policy condition and weights into the fake-account risk score. Disposable and invalid are strong enough to act on almost by themselves; webmail and accept-all are softer signals you weigh in rather than block on. The usual approach is to weigh them together rather than block on any one. A disposable address from a fingerprint that just created ten accounts is a far cleaner signal than the email property alone.

Rupt reads the email you send it. It never sends mail, and it doesn't verify the inbox for you: is_email_verified reflects what you've told Rupt about the user.