How the SMTP handshake works

SMTP verification follows the delivery conversation defined in RFC 5321, stopping just short of sending a message. Before any connection is made, the verifier checks syntax and looks up the domain’s MX records in DNS to find which servers accept its mail (see what is an MX record). It then:

  1. Opens a TCP connection to the highest-priority mail server on port 25 and waits for a 220 greeting.
  2. Identifies itself with EHLO (or HELO).
  3. Declares a sender with MAIL FROM.
  4. Declares the address being tested with RCPT TO.
  5. Reads the reply, then sends QUIT instead of DATA.

A simplified session looks like this:

S: 220 mx.example.com ESMTP ready
C: EHLO verifier.example.net
S: 250 mx.example.com
C: MAIL FROM:<probe@verifier.example.net>
S: 250 2.1.0 OK
C: RCPT TO:<jane.doe@example.com>
S: 250 2.1.5 OK            <- mailbox accepted
C: QUIT
S: 221 2.0.0 Bye

The reply to RCPT TO is the core signal. Because the session ends before DATA, the receiving server never gets a message body and the recipient sees nothing.

Reading the server’s reply

SMTP replies use a three-digit code, and most modern servers add an enhanced status code from RFC 3463 that gives more detail.

Reply Meaning Typical interpretation
250 2.1.5 Recipient OK Mailbox likely exists
550 5.1.1 Bad destination mailbox Mailbox does not exist
550 5.2.1 Mailbox disabled Account exists but is not accepting mail
452 4.2.2 / 552 5.2.2 Mailbox full Account exists but is over quota
450 / 451 Temporary failure Greylisting, throttling, or a transient problem
550 5.7.1 Delivery not authorized Policy or reputation block against the sender, not a verdict on the mailbox

The first digit matters most. A 2xx reply means accepted, 4xx means “try again later,” and 5xx means a permanent rejection. Treating a 4xx as invalid is a common mistake that inflates false-invalid rates. Equally, a 5xx that cites policy, spam, or a blocklisted IP says something about the verifier’s connection, not about whether the mailbox exists, so it should not be reported as undeliverable. Our bounce code reference lists the common codes and what they mean.

Why SMTP checks need dedicated infrastructure

Most cloud providers and residential ISPs block outbound port 25 by default to limit spam, so a typical application server cannot open these connections at all. Receiving servers also watch for probing behavior. Many connections from one IP in a short window can trigger rate limits, temporary deferrals, or blocklisting, which then makes every later result from that IP unreliable. Running SMTP checks at scale requires reputation-managed sending IPs with correct reverse DNS, sensible concurrency limits per destination, and retry scheduling that respects the receiving server.

Where SMTP verification alone falls short

An accepted RCPT TO is strong evidence, not proof. Several situations produce ambiguous results:

  • Catch-all domains accept every address at the RCPT TO stage, real or not. See what is a catch-all email domain.
  • Greylisting returns a temporary 4xx to unfamiliar senders and only accepts on a later retry. See what is greylisting.
  • Accept-then-bounce servers accept during the session and send a bounce later, often to resist address harvesting.
  • Disabled VRFY: RFC 5321 permits servers to turn off the VRFY command, so verifiers rely on RCPT TO instead.

This is why a complete pipeline combines SMTP with syntax checks, DNS and MX lookups, disposable and role-based detection, and catch-all analysis.

Practical steps

  • Verify new and imported lists before sending, and re-verify older segments periodically, since mailboxes are closed over time.
  • Retry 4xx results after a delay instead of discarding them. RFC 5321 suggests retry intervals of at least 30 minutes for real deliveries.
  • Keep “unknown” separate from “undeliverable” so you do not delete good contacts.
  • Suppress hard-bounced addresses immediately after every campaign.

Where RocketVerifier helps

RocketVerifier runs the full pipeline, including the SMTP handshake, from managed infrastructure and returns deliverable, undeliverable, risky, or unknown with a specific sub_status such as mailbox_not_found, greylisted, or blocked. Policy rejections are reported as unknown rather than undeliverable, and unknown results are never charged. You can test a single address with the free email verifier or see how each verdict is explained in explainable verification.