How do you check that an email address exists without sending an email?

By opening a conversation with the domain's mail server, the way a sender would, and stopping just before the message is transmitted. The server usually indicates whether the mailbox exists or not. No message is transmitted, the recipient receives nothing and is notified of nothing.

Is this verification one hundred percent reliable?

No, and no serious service claims that it is. About three quarters of addresses get a clear verdict. The rest are catch-all domains that accept everything, servers that refuse checks, or servers that answer after a deliberate delay. A good tool shows the reason rather than an invented percentage.

Why not simply send a test email?

Because sending to a nonexistent address produces a hard bounce. Above roughly two percent of bounces, mailbox providers degrade the reputation of the sending domain, and the messages that follow land in the junk folder, including those addressed to existing customers. A test email also reveals your approach before you have decided to write.

Does verification alert the person concerned?

No. The conversation with the server stops before the step where the content is sent. Nothing arrives in the person's mailbox, and nothing shows up in their junk folder.

Verify an email address without sending a message

Verifying an address means asking the recipient's mail server whether the mailbox exists, then hanging up before sending anything. The answer is the same one a real send would get, without the message and without the risk.

By Bertrand Menesguen, founder of Créacodeal · Updated : 2026-09-27

The principle

When a message goes out to [email protected], the sender's server first looks up which server receives mail for the domain, introduces itself, announces the sender, announces the recipient, then transmits the content. The receiving server accepts or refuses the recipient before the content is transmitted.

Verification stops at exactly that point: it obtains the acceptance or the refusal, then closes the conversation. Nothing is transmitted, nothing is delivered.

The four checks, from the cheapest to the most costly

Syntax. A malformed address is set aside immediately, with no network call.
The domain. The domain must exist and carry MX records. A domain with no MX receives no mail at all: every address built on it is unusable.
The mailbox. This is the dialogue described above, and it gives the verdict in the great majority of cases.
Public traces. When the server does not settle the matter, an address published somewhere on the web remains solid proof of existence.

The verdicts, and what they allow

Verdict	What it means	What to do
Valid	The server confirms that the mailbox exists, or the address is published on the web.	Write.
Likely	Catch-all domain, but the company's format is confirmed and the person is in post.	Write, accepting a low risk, and stay away from bulk sending.
Uncertain	Catch-all domain with no other clue.	Keep for a manual send, or look for another channel.
Nonexistent	The server explicitly refuses this mailbox.	Never write: a hard bounce is guaranteed.
Undetermined	The server did not answer clearly, or deferred its answer.	Try again later, the answer often changes.

The limits, stated plainly

Catch-all domains

Some domains accept every address, whether it exists or not. The server answers favorably to anything, which leaves the check silent. See catch-all domain. They account for roughly one address in five or six in France.

Servers that protect themselves

Some servers deliberately answer with a temporary response to a caller they do not know, others refuse any check that does not lead to a send. The verdict becomes undetermined, and a second pass a few hours later often settles it.

A snapshot, not a guarantee

An address that is valid today can disappear tomorrow when the person leaves. That is why verification has to be run again at the time of the campaign, rather than taken from a file built a year ago.

Why this matters for your deliverability

Mailbox providers judge a sender on its bounce rate. Above roughly two percent, the reputation of the domain degrades and every message that follows is filtered more harshly, including transactional messages and those going to existing customers. Cleaning a list before sending costs a few cents per address. Repairing a damaged reputation takes weeks.

What Veriphee does

Veriphee runs the four checks in sequence automatically, tries several spellings of the name when the first one does not answer, keeps the confirmed format for the next people at the same company, and shows the verdict in plain words. If no valid address is found for a company, half of the credit is returned. No message is sent at any point.

Frequently asked questions

How do you check that an email address exists without sending an email?: By opening a conversation with the domain's mail server, the way a sender would, and stopping just before the message is transmitted. The server usually indicates whether the mailbox exists or not. No message is transmitted, the recipient receives nothing and is notified of nothing.
Is this verification one hundred percent reliable?: No, and no serious service claims that it is. About three quarters of addresses get a clear verdict. The rest are catch-all domains that accept everything, servers that refuse checks, or servers that answer after a deliberate delay. A good tool shows the reason rather than an invented percentage.
Why not simply send a test email?: Because sending to a nonexistent address produces a hard bounce. Above roughly two percent of bounces, mailbox providers degrade the reputation of the sending domain, and the messages that follow land in the junk folder, including those addressed to existing customers. A test email also reveals your approach before you have decided to write.
Does verification alert the person concerned?: No. The conversation with the server stops before the step where the content is sent. Nothing arrives in the person's mailbox, and nothing shows up in their junk folder.