What Is SMTP Verification & How It Works
SMTP verification is the step that makes email verification worth paying for: the verifier connects to the recipient's own mail server and asks, in the server's own protocol, whether it would accept mail for that address. The server is the authority on its own mailboxes, so its answer is the closest thing to ground truth available without sending a message.
The conversation, step by step
SMTP is a text protocol, and a verification is a short, deliberately incomplete conversation:
- Find the mail servers — look up the domain's MX records to learn which servers accept its mail, in priority order.
- Connect — open a TCP connection, normally on port 25, and wait for the server's greeting.
- EHLO — identify the connecting host by name. Servers check this against reverse DNS, and a mismatch is a common reason for being refused.
- STARTTLS — upgrade the connection to TLS where the server offers it. Some servers reveal less, or refuse entirely, over an unencrypted connection.
- MAIL FROM — name a sender address. This must be a real, deliverable address on a domain with proper DNS, because many servers verify it before answering anything else.
- RCPT TO — name the address being checked. The reply to this command is the answer.
- QUIT — disconnect without sending DATA, so no message is ever transmitted.
Reading the reply codes
The numeric reply to RCPT TO carries the result:
- 250 — accepted. The server would take mail for this recipient.
- 550 / 551 / 553 — rejected. No such mailbox. This is a definite negative.
- 450 / 451 / 452 — temporary failure. Try again later — often greylisting, sometimes load. Treating this as a rejection is a false negative on a good address.
- 421 — service unavailable or the connection is being dropped, frequently rate limiting.
- 252 — the server will not confirm or deny the recipient but will accept the message. Explicitly ambiguous.
The quality of a verifier is largely determined by how it handles everything that is not 250 or 550. Reporting a greylisted address as invalid is the most common way to lose a real contact.
Why this is harder than it sounds
Receiving servers are defensive, and for good reason: a client that opens many connections, names many recipients and never delivers anything is behaving exactly like a directory harvest attack. The defences a verifier has to work with include:
- Greylisting — unknown senders are temporarily rejected on first contact and must return after a delay. A real mail server retries; spam software usually does not.
- Rate limiting — too many probes from one address get throttled or blocked, so the work has to be spread.
- Blocklists — an IP that behaves badly ends up on one, and then its answers get worse everywhere.
- Reverse DNS requirements — many servers refuse connections from IPs with no matching PTR record.
- Deliberate ambiguity — some providers never reveal whether a mailbox exists, as a privacy measure.
This is the main reason running your own checks from an office or cloud connection does not work well, and why providers invest in dedicated infrastructure, careful pacing and retry logic.
Where it cannot give a clear answer
On a catch-all domain the server accepts every recipient by configuration, so a 250 means only that the domain accepts mail — not that this mailbox exists. Some large providers also accept everything at the SMTP stage and decide later, which produces the same ambiguity.
A verifier that reports these as valid is overstating what it knows. The honest answer is accept-all, together with whatever secondary evidence is available.
Does it send email?
No. The conversation stops before the DATA command, which is the point at which a message body would be transmitted. Nothing is delivered, nothing appears in the recipient's mailbox, and the recipient has no way to observe it. Only the server's own logs record that a connection asked about an address.
Common questions
Does SMTP verification send an email to the recipient?
No. The check ends after the RCPT TO command and disconnects before any message content is sent, so nothing is delivered and the recipient cannot tell it happened.
Why do some servers accept every address?
They are configured as catch-all, accepting mail for any address at the domain and sorting it afterwards. Some providers also accept at the SMTP stage deliberately, to avoid revealing which mailboxes exist.
Can I run SMTP verification myself?
Technically yes, but most networks block outbound port 25, and servers commonly refuse connections from IPs without reverse DNS or sending history. Handling greylisting and rate limits correctly is also where most homemade checks produce wrong answers.