SMTP verification of an email address
SMTP verification means asking the server that hosts a mailbox whether it would accept a message for it, without ever sending one. The dialogue stops before the message is written. It is the only way to get an answer from the server on whether an address exists, and the only charged check on the platform.
- No message sent
- Dialogue on port 25
- Unknown, so not charged
The mechanism
What the probe actually does
The probe connects to the mail server announced by the domain, introduces itself, announces a sender address, then announces the address to verify. The server accepts or refuses that recipient. The probe then cuts the session, before the step that would transmit the content of the message.
-
Find the server
The mail records of the domain designate the server that receives its mail.
-
Open the session
The connection is made on port 25, the one used for mail between servers.
-
Announce the sender then the recipient
This is the step at which the server says whether it would accept the message.
-
Cut before sending
The session stops there. Nothing is written, nothing is transmitted, nothing arrives in the mailbox.
No message, ever
SMTP verification opens a session with the receiving server, announces the sender and the recipient, then stops before writing the message. No mail arrives in the mailbox being analysed, and the person whose address is verified receives nothing.
curl -X POST https://api.yeswecheck.fr/v2/email/validate \
-H "Authorization: Bearer $YESWECHECK_API_KEY" \
-H "Content-Type: application/json" \
-d '{"email":"[email protected]","smtp":true}' The answers
The six possible answers from the server
A receiving server does not only answer yes or no. It can accept, refuse, accept everything without distinction, delay, stay unreachable, or report a full mailbox. YesWeCheck tells each situation apart, returns the status that matches it, and only charges for those that make a verdict possible.
-
The server accepts the recipient and the domain does not accept everything.
valid valid address
One credit charged.
-
The server permanently refuses the recipient.
invalid invalid address
One credit charged.
-
The domain accepts every address: it is a catch-all domain.
unknown unknown
No credit charged.
-
The server answers with a delay, greylisting or equivalent.
unknown unknown
No credit charged.
-
The server stays unreachable or gives no usable answer.
unknown unknown
No credit charged.
-
The mailbox exists but it is full.
risky risky address
One credit charged.
The rule
The result of the probe prevails
When the probe actually talked to the server, its answer takes precedence over every other check. A risk signal can degrade a valid result to a risky result, never the reverse. A refusal by the server never becomes an acceptance because the other checks were favourable to the address.
When the SMTP probe actually talked to the server, its result prevails. The other checks can degrade a result to risky, they can never turn a refusal by the server into an acceptance.
Three signals degrade a valid result to a risky result, even when the server has confirmed the mailbox: a disposable domain, a malicious domain and a typosquatted domain. These addresses exist, they are deliverable, and they stay risky for a business reason rather than a technical one.
- Disposable domain
- Malicious domain
- Typosquatted domain
A refusal by the server stays a refusal
invalid the server refuses the recipient + every other check in the green valid never A risk signal can degrade
valid the server accepts the mailbox + disposable domain risky result returned The cost
What a probe costs
It is the only charged check on the platform, and it is only charged when it reaches a verdict. A verification that did not make a verdict possible costs nothing, whether the server delayed, stayed unreachable or accepted every address. The other nineteen checks never consume a credit.
- 1 credit when the probe reaches a verdict: valid, invalid or risky
- No credit when the server does not make a verdict possible: unknown
- 19 other checks run alongside the probe, never charged
- Adjustable timeout from 1,000 to 30,000 milliseconds per session
A credit is charged if and only if SMTP verification was requested and the final status is not "unknown".
The time granted to the session can be set between 1,000 and 30,000 milliseconds: you choose the balance between responsiveness and thoroughness that suits your journey.
Without SMTP verification, the response is immediate. With SMTP verification, it arrives as soon as the receiving server has answered, and the maximum time allowed is set on each call.
Questions
Frequently asked questions
- Does SMTP verification send a message to the address?
- No. SMTP verification opens a session with the receiving server, announces the sender and the recipient to it, then stops before writing the message. No mail arrives in the mailbox being analysed, and the person whose address is verified receives nothing.
- Why do some addresses stay unknown?
- Because the receiving server did not make a verdict possible: the domain accepts every address, the server applies a delay, or it stays unreachable. Rather than guess, YesWeCheck then returns an unknown status, and no credit is charged.
- Can another check contradict the result of the SMTP probe?
- When the SMTP probe actually talked to the server, its result prevails. The other checks can degrade a result to risky, they can never turn a refusal by the server into an acceptance.
- Is SMTP verification always charged?
- A credit is charged if and only if SMTP verification was requested and the final status is not unknown. A probe that did not make a verdict possible costs nothing, in real time as in batch processing.
Going further
SMTP verification works the same way in real time and in batch processing, with the same billing rules. Batch processing adds a second pass, which probes the addresses left unknown again later and catches up with the servers that delayed during the first.