Email verification combines several methods, each proving a different thing: syntax and RFC checks (is it a valid address), MX lookup (can the domain receive mail), SMTP mailbox probing (does the mailbox answer), catch-all and disposable detection (can the result be trusted), and pattern matching (is it the likely format). No single method is enough; a real verifier runs them in sequence and reports a combined verdict. The honest limit is that catch-all servers and blocking gateways leave about 27% of business addresses unconfirmable, which should be reported as unknown, not valid.
| Method | What it checks | Proves | Limit |
|---|---|---|---|
| Syntax / RFC | The address is correctly formed | It could exist | A valid shape is not a live mailbox |
| MX lookup | The domain publishes a mail server | The domain can receive mail | Says nothing about the specific mailbox |
| SMTP probe | The server accepts the mailbox | The mailbox exists | Catch-all servers accept everything |
| Catch-all detection | Whether the server accepts any address | Whether the SMTP result can be trusted | Catch-all cannot be resolved further |
| Disposable detection | The domain against a throwaway list | The address is temporary | New throwaway domains appear daily |
| Pattern matching | The address against the company format | The address is plausible | Plausible is not confirmed |
| API / real-time | All of the above at the point of entry | The address is good before it is stored | Needs integration into the form |
Each method answers a narrower question than people assume. Syntax says the address is shaped right. MX says the domain can receive mail. Only the SMTP probe speaks to the actual mailbox, and even that is defeated by a catch-all server that accepts every address. A tool that stops at syntax and MX will pass millions of dead mailboxes; a tool that runs the SMTP probe but hides catch-all behind "valid" will pass guesses.
A real verifier runs them in order, stops when one gives a definitive no, and combines the rest into a single verdict: valid, invalid, catch-all, unknown, disposable or role. The value is in the sequencing and in reporting the uncertain cases honestly.
There is a hard limit no method can pass. Catch-all servers accept every address, so the mailbox probe cannot distinguish a real address from a guess, and some corporate and ISP gateways refuse the probe entirely. Across the business addresses we verify, 27% come back unknown for these reasons and about one in forty is catch-all. Any tool claiming near-100% accuracy with no unknowns is folding those cases into "valid". Honest accuracy means labelling them, and not charging for the unknowns, which is what we do.
Syntax and RFC checks, MX record lookup, SMTP mailbox probing, catch-all and disposable detection, pattern matching, and real-time API checks. Each proves a different thing, so a real verifier runs them in sequence and combines them into one verdict rather than relying on any single method.
The SMTP mailbox probe is the only method that speaks to whether the specific mailbox exists, so it is the most decisive, but it is defeated by catch-all servers. Accuracy comes from combining it with MX, syntax, catch-all and disposable detection, and from honestly labelling the cases that cannot be confirmed rather than guessing.
Yes. The SMTP probe opens a connection to the receiving server and asks whether the mailbox exists, then closes before any message is sent. Every standard verification method is passive; the recipient is never contacted and your sender reputation is never touched.
Because the receiving gateway refused the mailbox probe, or the domain is catch-all and accepts everything. Neither can be resolved by any method. Across business domains, 27% return unknown this way. An honest tool reports them as unknown; a tool that reports them as valid is guessing.
In practice they are used interchangeably, but validation often refers to the lighter checks (syntax, MX) and verification to the full process including the SMTP mailbox probe. A real product does both; checking only syntax and MX leaves most dead mailboxes undetected.
Syntax, MX, SMTP, catch-all and disposable detection in one pass, with honest unknowns. 100 checks a month at no cost.
Ayoub built BounceZero's 5-stage validation pipeline, its dedicated sending infrastructure, and the high-availability database behind every verification. Previously built high-volume email delivery systems. Trained at 1337 Benguerir (École 42 network, 2019). Open-source: bgp_analyzer.
Single-email checks, no signup required
Check if an address exists before you send
Predict a bounce before sending
The cost of a check vs the cost of a bounce
Check if any email address is real
Instant real-time single address check
Verify a whole list up to your free monthly credits
Continue through related topics