One verification is twenty-eight measurements behind a single verdict. All of them run on every plan, including Free.
Catching bad input before it costs a lookup.
RFC-correct parsing of the local part and domain, quoted strings and internationalized mail included.
Did-you-mean suggestions for common misspellings of major providers, returned as data. We never silently change the address you submitted.
Keyboard-walk and random-string local parts flagged before they read as real people.
What the domain itself tells us.
Mail exchangers resolved and ranked, including implicit MX when no MX exists.
Domains that declare they accept no mail (RFC 7505) are invalid without a connection.
110,000+ throwaway providers, refreshed from two maintained lists.
info@, sales@, admin@ and friends, flagged so you can mail them differently.
Gmail, Outlook and company identified, since free mailboxes behave differently in lists.
Registrant placeholders and expired services identified by their DNS fingerprints.
Dead DNS confirmed against an independent resolver before a domain is called invalid.
Talking to the server that actually answers for the address.
A real conversation with the responsible mail server, banners and reply codes classified per RFC 3463.
Reply quirks read the way each operator means them: IronPort, Exim, GoDaddy, qmail and more.
Dedicated strategy with target-first probing and per-domain behavior profiles.
Consumer accepts are confirmed through Microsoft's GetCredentialType API before a valid verdict.
Verified through Yahoo's account-creation validation endpoint with browser-accurate TLS.
Provider-specific response classes for iCloud, Proton and Zoho mail.
United Internet (GMX, WEB.DE, 1&1), T-Online, Orange, SFR and free.fr egress handled per their own rules.
Mimecast and Proofpoint front-ends recognized and answered accordingly.
Turning measurements into a number you can automate against.
Accept-all domains are measured with control probes; without proof of existence the verdict caps at risky.
SPF, DKIM, DMARC and SpamAssassin-style content signals where they apply.
Patterns associated with recycled and pristine traps weighted into the score.
Blocklist presence for the domain and its mail servers.
A small but real signal that a human uses the address.
Likely-mistyped addresses scored down even when the typo happens to exist.
Same 28-check suite, no card required.