How catch-all domains get around email verification
A catch-all domain passes email verification on every new address. Why verified doesn't mean new, which basic fixes break, and who they wrongly block.
Catch-all domains get around email verification because every made-up address on them still receives the code. It's Wednesday morning, the day after your launch, and you're looking at the best numbers you've ever had. The $2,400 newsletter sponsorship you agonized over brought in 1,240 signups, and 1,116 of them confirmed their email. That's 90%. You've already sent your co-founder the screenshot.
At 9:24 a customer on your $79 plan writes in: "Every run has failed since before 7. Are you down?" You open your model provider's dashboard. The $3,000 monthly limit you set was reached at 6:52 a.m., and since then your product has been returning errors to everyone. You raise the limit from your phone at 9:31, which means paying customers were locked out for 2 hours and 39 minutes on the morning you thought was your best ever.
Then you go looking for whatever ate the budget. You sort the new signups by email domain, and three of the biggest belong to no company you can find anywhere. One has 186 accounts, the next 142, the next 97. Every address is different, every one is verified, and every one spent its full 500 free credits overnight. This is catch-all domain signup abuse, and your email verification did its entire job while it happened.
This is a scenario, so check the sums yourself:
- 186 + 142 + 97 = 425 accounts
- 425 accounts × $3.20 of model cost per 500 credits = $1,360
- $1,640 already spent this month + $1,360 = $3,000, your limit
- 3 customers refunded for the outage × $79 = $237
- $1,360 + $237 = $1,597 spent on people who were never going to pay you
If your company spends $12,000 a month, that's $400 a day, so $1,597 is 4 days of runway. The launch looks worse too. You told your co-founder you'd paid $2.15 per verified signup ($2,400 ÷ 1,116). Take out the 425 and it's $3.47 ($2,400 ÷ 691).
At 10:05 you decide to block every domain like these. You run yesterday's verified list through an email validation tool and ask which addresses sit on domains that accept all mail. It returns 487. The first 425 you expected. The other 62 are a 12-person architecture firm, a consultant who runs their whole business from one domain, and a design studio that set things up this way years ago so no client email would ever get lost.
Say your trial converts at 8% and those people would pay $79 a month. 62 × 8% is about 5 customers, and 5 × $79 × 12 months = $4,740 a year. If you block them, they'll see an error and close the tab, and not one of them will write to tell you why. So you have two ways to lose. Leave it open and pay $1,360 again next launch, or shut it and lose $4,740 a year that no report will ever show you.
What a catch-all domain is
A catch-all domain is a domain whose mail server accepts mail for any address at that domain, whether or not a mailbox with that name exists. Mail to anything@example.com gets delivered to one inbox, or to whoever the owner chose.
It's an ordinary mail setting, and most people who use it have dull reasons. A misspelled name still reaches someone. A former employee's address still gets read. Email validation tools often label these domains "accept-all", because the server says yes to every address and the tool can't tell whether any particular mailbox exists.
That's what makes it matter for your signup form. On most domains, one person has one address. On a catch-all domain, whoever holds the domain can receive mail at as many addresses as they care to write. Say a domain costs $12 a year. The 186 accounts on that first domain took $595.20 of credits (186 × $3.20), about 50 times what the domain cost.
Why a verified email doesn't mean a new person
Email verification asks one question: can someone read mail sent to this address? On a catch-all domain the honest answer is yes, every single time.
Nothing in that exchange failed. Your app sent a real code to a real address, and someone read it. What your app then did with that result is where the money went, because in most codebases "verified" quietly turns into "new person, give them the credits".
Blocking the domain feels like the fix. It holds for exactly as long as the next signup uses that same domain, though. And "accepts all mail" describes how a server is configured, so a rule built on it can't tell the architecture firm from the 186.
The email fixes people try first, and where they stop working
Keep email verification. It catches typos and dead addresses. It misses any address that can receive mail, it occasionally turns away someone whose code landed in spam, and its upkeep is deliverability work that never quite ends.
A unique-email rule catches the same address used twice and wrongly blocks almost nobody. It's cheap, and all 186 different addresses pass it.
Personal Gmail ignores dots, and mail to jane+notes@gmail.com lands with Jane, so rules for this catch one Gmail inbox written several ways. They do nothing for a catch-all domain, where each address really is different.
Disposable-domain blocks catch known throwaway services. A privately held domain isn't a throwaway service, so it usually isn't on the list. The list also hits people using privacy forwarding, and someone has to keep updating it forever.
Blocking every catch-all domain would have caught all 425. It would also have turned away the 62, and you'd be paying for a lookup on every signup to do it.
A cap per business domain does better. Allow 5 free grants per domain per day, and yesterday's three domains get 15 accounts, or $48 of credits (15 × $3.20) in place of $1,360. That's a real improvement, but it only ever counts one domain at a time. It also stops a real company rolling out to its whole team on the day they decide to buy, so those teams need a way to reach you.
The other basics
| Fix | What it catches | What it misses | Who it wrongly blocks, and the upkeep |
|---|---|---|---|
| CAPTCHA | Simple scripts | People signing up by hand | Screen reader users; fees and lost signups |
| Phone requirement | Casual second accounts | Anyone with another number | People without a reliable mobile; a charge per text |
| Card requirement | Casual second accounts | Anyone with another card | People who want to try first; fewer signups |
| One signup per IP | Repeats on one connection | Any other network | Offices and households; exceptions forever |
| VPN blocks | Known VPN ranges | Everything off your list | Staff on company VPNs; a second list to maintain |
| Rate limits | Bursts from one source | Slow, spread-out signups | Busy shared networks; tuning per route |
| Usage caps | One account overrunning | Each new account's fresh cap | Anyone at the cap, usually your most active users |
| Shorter trials or smaller grants | Some cost per account | More accounts make up the difference | Every honest user gets less |
| Manual review | What a person can spot | Anything after the credits are spent | Whoever looks odd to a tired reviewer |
1,116 verified signups at one minute each is 18.6 hours of review, and on Wednesday morning the credits were gone before anyone started.
What to do about catch-all domains
None of those fixes is reliable alone, and stacking them adds friction for everyone you wanted. This order gives you the most protection for the least damage.
Change what "verified" unlocks. Treat it as "we can reach this address". Don't let that one flag release anything that costs you money. Hold expensive work until an account has done something real.
Treat accept-all as a fact about mail. Don't deny anyone on that label alone, because real businesses carry it.
Cap free grants per business domain, and put a "talk to us" path on the limit so a real team gets through in one reply.
Give free usage its own budget, with a hard spending cap. That 6:52 a.m. outage happened because free and paid shared one limit.
Report launches on people who used the product. Verified signups inflate the number you're making decisions with.
Write the no before you need it. Some people you turn away will be real, so say it kindly and give them a way to reply.
Do all six and the next launch will hurt less. You'll also have a cap to tune, a support path to staff and one question still open: has the person behind this brand-new verified address already had this benefit? Everything above looks at an address or a domain. That question is about a person, and the real people caught in the middle make a wrong answer expensive.
When verified isn't the answer you need
You don't have to build your own abuse prevention system, and this is the point where most teams should stop trying. AbuseZero checks mailbox aliases, disposable domains and mail routing. Your domain cap counts one domain. AbuseZero recognizes repeat devices across new accounts, even when email addresses keep changing, so creating a fresh account isn't enough to erase previous activity. You get multi-layer AI risk analysis that examines 140,000+ signals, detects linked claims and flags abuse that single checks miss. There's also a proprietary algorithm, with AI that checks the claim again.
Your server makes one API call before the credits go out and gets a clear allow or deny, under rules you control. If the 62 are what keep you up at night, Observe mode shows you what would be denied before anyone is blocked. The quickstart gets your first benefit protected.
Frequently asked questions
Should I block catch-all domains at signup?
Usually not outright. A blanket block turns away real companies with no explanation. A cap on free grants per domain, with a way to reach you, costs far less.
Does a verified email prove someone is a new user?
No. It proves someone can read mail at that address. Whether this claim looks like a repeat is a separate question, and it's the one AbuseZero's evaluation is there to answer.
Are all accounts on one domain suspicious?
No. Colleagues share a domain by design, so a returning customer or a colleague can look a lot like a duplicate. Write down what you mean to give each person first.
Save your free credits for new people
Start with Builder
Pay as you go. Protect your next launch.
See pricing
Builder, Pro, Scale and Enterprise.
Talk to us about Enterprise
High volumes and custom terms.