How to check if an email address exists (without sending an email)
Check if an email address exists without sending anything: the SMTP conversation step by step, what replies mean, catch-all, greylisting and honest unknowns.
To check if an email address exists without sending an email, ask the domain’s mail server. Mail servers decide whether to accept a recipient before any message is transferred, so you can open a normal SMTP conversation, name the address, read the reply and hang up. Do it carefully and it tells you whether [email protected] is a real mailbox, a typo, a full inbox or a domain that accepts anything.
If you just want the answer for one address, paste it into the free email verifier. The rest of this post explains what happens behind that button, and where the answer can and can’t be trusted.
Step 1: rule out the cheap failures
Before talking to any mail server, three checks catch most bad addresses:
- Syntax. One
@, a valid domain, no illegal characters.jane@acmeandjane.acme.iostop here. - Typos.
[email protected]is valid syntax at a domain nobody uses. A typo check suggestsgmail.cominstead. - The domain and its MX records. If the domain doesn’t exist, or publishes no mail servers, nothing can be delivered. A null MX (
MX 0 ., from RFC 7505) is the domain saying outright that it accepts no mail. With no MX at all, RFC 5321 lets a sender fall back to the domain’s address record; if there isn’t one either, the address is dead.
You can see any domain’s mail servers with the MX lookup. It is also worth knowing what kind of address it is: a disposable inbox will “exist” for an hour and then vanish, which is why classification comes first.
Step 2: ask the mail server
Connect to the domain’s primary mail server (the MX with the lowest priority number) on port 25 and talk SMTP, the same protocol every sending server uses:
S: 220 mx.acme.io ESMTP
C: EHLO verifier.example.com
S: 250-mx.acme.io
S: 250 PIPELINING
C: MAIL FROM:<[email protected]>
S: 250 2.1.0 OK
C: RCPT TO:<[email protected]>
S: 250 2.1.5 OK <- the server accepts this recipient
C: RCPT TO:<[email protected]>
S: 550 5.1.1 User unknown <- and refuses a made-up one
C: QUIT
S: 221 Bye
The important line is the reply to RCPT TO. To deliver mail at all, a server has to say whether it will take a message for that recipient, and most servers that know their users say no to unknown ones right there. The conversation ends with QUIT before DATA, the command that would start a message. Nothing is sent and nothing lands in anyone’s inbox.
The older VRFY command was designed for exactly this question, but most servers switch it off, so verification uses RCPT TO.
Step 3: test a made-up address
A 250 for jane@ is only half an answer. Some domains accept mail for every address, real or not, and sort it out later: they are called catch-all or accept-all domains. To tell them apart, ask about a random address at the same domain in the same conversation, as in the example above.
jane@ |
Random address | What it means |
|---|---|---|
| Accepted | Refused | The mailbox exists. |
| Accepted | Accepted | Catch-all domain: the 250 proves nothing. |
| Refused | (not needed) | No such mailbox. |
Ask about the real address first. Some servers hang up after one unknown recipient to slow down address guessing, and you want the answer that matters before that happens. What is a catch-all email domain? covers what to do with those addresses.
Step 4: read the reply correctly
This is where most homemade checkers go wrong. SMTP replies come in three useful families:
- 2xx: accepted.
- 4xx: “not now, try again later”. Greylisting, a full queue, a rate limit.
- 5xx: a permanent refusal. Usually “no such mailbox”, but not always.
A 550 is used both for “this user doesn’t exist” and for “we don’t accept connections from your IP”. Treat a block as a missing mailbox and you’ll tell your users that real colleagues’ addresses are invalid. Separate them with the enhanced status code and the wording:
| Reply | Means |
|---|---|
550 5.1.1 User unknown |
No such mailbox |
550 5.2.1 ... is disabled |
Mailbox exists but is disabled |
452 4.2.2 ... over quota |
Mailbox exists but is full |
550 5.7.1 ... blocked using zen.spamhaus.org |
Your check was blocked, the mailbox is unknown |
451 4.7.1 Greylisting in action |
Ask again in a few minutes |
Providers also have their own dialects: Microsoft 365 rejects unknown recipients with a 5.4.1 that says “Access denied”, which reads like a block but isn’t. Email bounce codes explained has the full list we handle.
Why “unknown” is an honest answer
Sometimes the server doesn’t give a usable answer:
- Greylisting. Servers that defer unfamiliar senders with a
4xx(described in RFC 6647) expect you to come back in a few minutes. The right response is to retry later, not to guess. - Blocks and reputation. Mail servers rate-limit or block IPs that open many conversations without sending mail. A refusal at the greeting, at
MAIL FROMor with a block-list reason atRCPT TOtells you about the checker, not the mailbox. - Timeouts and unreachable servers. Silence is not a “no”.
- Backup MX servers. Secondary mail servers often accept every recipient because they can’t see the user directory. Only fall back to them when the primary can’t be reached, and trust their answer less.
A checker that turns these into “invalid” will make you throw away good addresses. A checker that turns them into “valid” will let bounces through. Saying unknown, with the reason, is the accurate answer.
Doing it at scale without breaking things
Running SMTP checks yourself is possible, but a few practical points decide whether it keeps working:
- Outbound port 25 is blocked by default on many cloud providers.
- The connecting IP needs a sensible identity: the
EHLOname should match the IP’s reverse DNS, or many servers refuse the conversation. - Pace your checks per provider. A burst of recipient checks to Gmail or Microsoft looks like address harvesting, and a blocked IP turns every later answer into “unknown”.
- Cache answers. A mailbox rarely changes from one minute to the next, and asking twice costs reputation.
- Don’t store the address just to cache it. A keyed hash is enough.
The isBusinessEmail email verifier
The free email verifier and the /v1/verify API endpoint do everything above for one address at a time:
- syntax, typo, domain, MX and null-MX checks, plus the address’s category (disposable inboxes are reported as risky and not probed);
- an SMTP conversation with the primary MX that stops before
DATA, then the catch-all test with a random address; - replies read with the enhanced status code and provider quirks, so a block is
unknown(probe_blocked), neverundeliverable; - greylisted checks retried in the background after about 90 seconds and 5 minutes, so asking again shortly afterwards gets the real answer;
- results cached for up to 24 hours under a keyed hash of the address, which is never stored or logged.
Every answer is one of four statuses: deliverable, undeliverable, risky (catch-all, full mailbox, disposable) or unknown, with a reason code and a confidence. The page allows 20 free verifications a day; with a free API key you can call it from your own code, 100 a day on the free tier and up to 10,000 a day on paid tiers. You can try the API without a key on test addresses such as [email protected].
The one test that is always right
An SMTP check tells you whether a server accepts mail for an address today. It can’t tell you whether a person reads it, or whether the person typing it into your form owns it. For sign-ups, the definitive check is still a confirmation email with a link. Use SMTP verification where a confirmation email isn’t possible: addresses you imported, collected on paper or typed in for someone else, and lists you are about to send to.
Frequently asked questions
Can you check if an email address exists without sending an email?
Yes, for most domains. You connect to the domain's mail server, start a normal SMTP conversation, give the address in RCPT TO and read the reply, then disconnect before any message is sent. Catch-all domains, greylisting and servers that refuse checks make some answers uncertain.
Is it legal to check whether an email address exists?
Asking a mail server whether it accepts a recipient is part of normal email delivery, but the address is still personal data. Check addresses you have a reason to check, such as one a person gave you, and mention verification in your privacy notice. Never use it to guess addresses.
Why does an email checker say 'unknown'?
Because the mail server didn't give an answer that can be trusted: it greylisted the check, timed out, or refused to answer because of a block list or policy. None of those say anything about the mailbox, so a careful checker reports unknown instead of invalid.
Does the person get notified when I check their email address?
No. No message is sent, so nothing arrives in their inbox. The receiving mail server does see the question and may log it.
What does the VRFY command do?
VRFY was designed to ask a server whether a mailbox exists, but most servers disable it or answer every query the same way. Verification tools use RCPT TO instead, because servers have to answer it truthfully to deliver mail.
How accurate is SMTP email verification?
For domains that reject unknown recipients during the conversation, which includes Gmail, Google Workspace, Microsoft 365 and Outlook.com, the reply is the same one a real message would get. Accuracy drops on catch-all domains and when the check is blocked, which is why those cases should be reported separately.