MailFloo › Catch-all email verification: how to resolve accept-all addresses

Catch-all email verification: how to resolve accept-all addresses

Most verifiers detect a catch-all domain and stop at “risky”. This is what a catch-all actually is, why the server's reply tells you nothing, and the four signals that resolve the address anyway.

Updated 18 October 2026·8 min read

What a catch-all domain actually is

Every mail server has a rule for addresses it doesn't recognise. Most say no: the SMTP reply is a 550 5.1.1 User unknown and the message bounces. A catch-all domain (Microsoft calls it accept-all, some admins say wildcard) is configured the other way: it says yes to everything at that domain and sorts it out internally, forwarding to a shared mailbox, dropping it silently, or quarantining it.

Companies set this up for sensible reasons: so a typo in a customer's address still arrives, so mail to departed staff can be redirected, so the sales inbox catches sales@, sale@ and salez@. Secure email gateways (Proofpoint, Mimecast, Barracuda) also behave this way by default, because they accept first and filter later. That is why catch-all is overwhelmingly a business phenomenon: in the lists we see, roughly a third to a half of B2B domains accept everything, while consumer providers like Gmail and Yahoo almost never do.

Why standard verification stops at "risky"

Email verification works by asking the recipient's mail server a question before any message is sent. The verifier connects on port 25, introduces itself, names a sender, then names the recipient with RCPT TO. A normal server answers honestly:

→ RCPT TO:<john.smith@acme.com>
250 2.1.5 Recipient ok
→ RCPT TO:<this-is-made-up-7x9q@acme.com>
550 5.1.1 User unknown

A catch-all server answers 250 to both. The real mailbox and the invented one look identical, so the server's reply carries no information about the individual address. Every verifier on the market detects this the same way, by sending an address that cannot exist and seeing whether it is accepted, and at that point most of them stop. The address is labelled catch-all, accept-all, risky or unknown, and the decision is handed back to you.

Detection is not resolution. Knowing a domain is catch-all tells you nothing about whether anna.henschel@ is a real person there. Detection is a domain-level fact; resolution is an address-level verdict. Most tools sell the first and call it the second.

Why it matters more than it looks

If a third of your business addresses come back "risky", you have three bad options. Send to all of them and accept the bounces, which damages the sending domain's reputation, the one asset a cold-email operation cannot buy back. Drop all of them and lose a third of your pipeline, most of which was perfectly good. Or pay a second vendor to "deep verify" the leftovers, which usually means they send a real email and watch for a bounce, the exact thing verification was supposed to avoid.

The grey zone is also where the money is. Enterprise domains, regulated industries and anything behind a security gateway skew heavily toward catch-all. Those are the accounts a B2B list is built to reach.

How MailFloo resolves a catch-all address

Once the probe pair confirms a domain accepts anything, the server's own reply is set aside as uninformative and four independent signals take over. Each contributes to a 0–100 score and a one-sentence reason, and the final label is Catch-all · likely, plain Catch-all, or (when the evidence is strong enough) Valid or Invalid.

1 · Naming pattern

Organisations assign addresses by rule: first.last@, flast@, first@. MailFloo collects the addresses already confirmed on a domain (from earlier verifications and from recorded deliveries) and derives the pattern. An address that follows the dominant pattern of fourteen confirmed colleagues is very likely real; one that follows no pattern at all is not. This is the single strongest signal for mid-sized companies, and it gets stronger the more of a domain you have seen.

2 · The server behind the gateway

A domain fronted by Proofpoint or Mimecast accepts everything at the gateway, but the organisation's actual mail server behind it often does not. Where a domain publishes a lower-priority MX that reaches that server directly, MailFloo asks it the same question. A 550 from the real server is definitive even when the gateway said yes.

3 · Delivery history

When you record what actually happened to a message (delivered or bounced), that fact outranks any server reply for that address, and it informs every other address on the domain. Three recorded deliveries to a catch-all domain with no bounces raises the whole domain's trust; a domain that accepts cheerfully and then bounces twice gets every future "valid" downgraded to risky. You can record outcomes one at a time, paste a list, upload your mailer's bounce report, or point an event webhook (SendGrid, Mailgun, Brevo, Postmark, SES) at MailFloo so it learns without anyone doing anything.

4 · Provider behaviour

Microsoft 365 is a special case worth naming. Tenants with directory-based edge blocking reject unknown addresses with 550 5.4.1: a definitive no that some verifiers misread as a server error. MailFloo treats it as the clean rejection it is. Similarly, the ratio of acceptances to recorded deliveries is tracked per provider, so an acceptance from a provider that has proven unreliable is worth less than one from a provider that has not.

What you get back

LabelMeaningWhat to do
ValidMailbox confirmed by its own server, or confirmed by recorded delivery.Send.
Catch-all · likelyDomain accepts anything, but the address matches the naming pattern and/or the domain has a delivery record. Score typically 60–80.Send; treat like a valid address with a slightly higher bounce allowance.
Catch-allDomain accepts anything and no further signal is available yet. Score typically 35–50.Send in a small, monitored batch first; record the outcomes and the rest of the domain resolves.
InvalidRejected by the server, by the server behind the gateway, or a real message to it bounced.Remove.

Every row also carries the sentence behind the verdict: "Follows the naming pattern of 14 confirmed addresses on this domain", "Gateway accepts all; the organisation's own server confirmed it", "A real message bounced 3 days ago", so a person can judge the call, not just read a label.

What catch-all resolution cannot do

It is worth being plain about limits, because the market is not. No method short of sending mail can be certain about a single address on a catch-all domain you have never seen before and have no delivery history for. A naming-pattern match is strong evidence, not proof; a gateway with no reachable server behind it gives nothing; a brand-new domain with one address has no pattern to match. MailFloo's answer in those cases is the honest one (Catch-all, score 45) and the practical advice is the same every verifier should give: send a small batch, record what happens, and let the loop do its work. Within ten minutes of recording outcomes, the rest of that domain re-scores.

Questions

What is a catch-all email domain?

A domain whose mail server accepts mail for any address, whether or not the mailbox exists. Instead of rejecting unknown recipients it routes them to a shared mailbox, drops them, or quarantines them. Microsoft calls the setting accept-all.

Are catch-all and accept-all the same thing?

Yes. Catch-all is the common term; accept-all is Microsoft's name for the same configuration; wildcard is older admin jargon. They all mean the server says yes to every address at the domain.

Why do verifiers mark catch-all emails as risky?

Because the server's reply carries no information: it accepts a real mailbox and an invented one identically. A verifier that relies only on the SMTP reply has nothing left to go on, so it reports the domain-level fact and labels the address risky or unknown.

Can a catch-all email address be verified without sending an email?

Partly, and much more than most tools attempt. MailFloo uses the naming pattern of addresses already confirmed on the domain, the organisation's real mail server behind a filtering gateway, recorded delivery outcomes, and per-provider behaviour. Together they resolve most catch-all addresses to a likelihood and a score. No method is certain for a single address on a never-before-seen domain.

Is it safe to email catch-all addresses?

It is safe when the address resolves to likely or valid. For unresolved catch-all addresses, send a small monitored batch first and record the results; the bounce rate of that batch tells you how to treat the rest of the domain, and MailFloo re-scores the domain from what you record.

Does MailFloo send a test email to resolve catch-all addresses?

No. Every signal is gathered from the SMTP conversation, DNS, previously confirmed addresses and outcomes you record. Nobody receives anything during verification.