How many domains show a Stripe verification record?

92,560 domains publish a stripe-verification TXT record at the apex, counted in the September 2026 census, and that counts published DNS records, which is a different thing from counting websites that take payments through Stripe. It comes to 0.06% of txt_present, the 167,692,006 domains where the census observed at least one apex TXT record, and 0.21% of token_present, the 44,285,266 domains carrying at least one verification-style token. The census saw the string in 381 TLDs.

You came here because someone wants a deployment figure for a slide and the only numbers on offer are the vendor’s. This one is different in kind. It is one exact string, read from public DNS on a stated date, with its denominator printed next to it, and anyone with a resolver can check a sample of it, which is why you can cite it and also why it can’t tell you more than it does.

Where the number comes from

Each month the census reads the apex TXT records of each domain it can resolve. Most of those records are SPF, DMARC and DKIM material, or something carrying a v= version tag. The rest are parsed for the shape a verification token takes: a string like <token>=<value> or <token>:<value>, where the part before the separator names whoever asked for the record.

stripe-verification is one of those token names. A domain counts once, however many copies of the string it carries and whatever value follows the equals sign. The census doesn’t read anything into the value and doesn’t check anything with Stripe. When it sees the record and the name matches, that TLD’s tally goes up by one.

That’s the whole method. It’s narrow on purpose, so that anyone can reproduce it, and in a memo that counts for more than a bigger number with no provenance.

What the count measures, and what it can’t

The figure counts domains that have published the string stripe-verification= at the apex. It doesn’t count Stripe accounts or checkouts, and it can’t tell you which sites accept a card through Stripe, because DNS shows none of that.

A domain can take payments with no token at all: an integration that never asked for a DNS record leaves nothing here, so those domains are invisible to this count. The gap runs the other way too, because a token can outlive the integration that created it: DNS records are rarely tidied when a feature is switched off, so some of the 92,560 are stale. You can’t tell from the outside which are which.

So treat the figure as a firm floor on published records, since the string is either in the zone or it isn’t, and don’t read it as the install base of a product. If the question you were asked is “how many websites use Stripe”, public DNS can’t answer it, and the count above is the nearest thing it can answer.

The counting rule leans one way

The census counts a token only where at least three domains in the same TLD carry it. A string that appears on one or two domains in a small TLD isn’t included in that TLD’s sum, which keeps typos and one-off oddities out of the table.

For a token as widespread as this one the rule matters at the edges, in TLDs where the string is rare. If anything, the rule makes the figure smaller. If you’re going to be wrong about this number, you’ll be wrong in the direction of understating it, and that’s a useful thing to be able to say when a figure is challenged.

Two denominators, and which one to quote

Which base you quote depends on the sentence you’re putting the figure in.

Of the 167,692,006 domains with at least one apex TXT record, 0.06% carry the token. Quote this one when the sentence is about domains that publish anything at all in TXT, which is the broadest population the count can sit inside.

Of the 44,285,266 domains carrying at least one verification-style token, 0.21% carry this one. Quote this when the sentence is about domains that have proven ownership to some third party, since that’s the population the token belongs to. Those 44,285,266 are themselves 26.41% of txt_present.

There’s no share of all registered domains here, because the census doesn’t publish one for tokens, and dividing by a different universe would give you a percentage nobody can reproduce from the source. Whichever base fits your sentence, name it in the sentence and don’t rework the arithmetic.

What the record looks like in your zone

If you own a domain and want to know if it’s inside the 92,560, one query settles it. On a Mac or Linux machine:

dig TXT example.com +short

On Windows:

nslookup -type=TXT example.com

A domain that’s in the count returns a line shaped like this, alongside whatever else lives at the apex:

"stripe-verification=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
"v=spf1 mx -all"

The first line is the one the census matches. The value after the equals sign is whatever Stripe issued at verification, and the census doesn’t inspect it. If you see the record and no longer know why it’s there, that’s the stale case from earlier, and you’ve found one of the reasons the count is a floor.

Run the free scan on your domain and it will read the same apex TXT set, along with the rest of the records the census grades, so you can see what the census saw.

How to cite it without overreaching

This sentence will survive a fact check:

“As of 5 September 2026, 92,560 domains published a stripe-verification TXT record at the apex, 0.06% of the 167,692,006 domains with at least one apex TXT record, according to the September 2026 edition of the defaults.exposed census.”

It states the date and names the base, and you can find each claim in it on the sheet. If your slide is about verified domains instead of TXT publishers, use the token_present base.

The line that will get you into trouble is “92,560 websites use Stripe”. It turns a DNS record into a product install and it turns domains into websites, when a domain can front several sites or none, so it overstates on both counts. If a reader challenges it, you’ll have no ground to stand on. Keep the word “record” in the sentence and you’ll be fine.

Don’t quote a change from an earlier month either. The census universe is rebuilt each edition and the source lists differ, so a raw difference between two editions mixes movement in the token with movement in the population. Cite the edition and the count as they stand, and when the next edition comes out it replaces them.

What else the same query shows you

The dig output above returned two lines, and only one of them was the point of this piece. The other was an SPF record, and it’s there because the apex TXT set is shared ground. Whatever verification tokens you’ve collected sit next to the records that decide whether mail claiming to be from your domain is accepted.

That’s worth a look while you have the terminal open. How SPF works and what the -all at the end means is a short read. DMARC lives one label down, at _dmarc.example.com, so it won’t show in the apex query, and it’s the record that tells receivers what to do when SPF fails. If your zone has grown a long tail of records nobody remembers adding, the DNS piece covers what each type is for and how to read a full zone without guessing.

None of that changes the count. All the token tells you is that the domain verified itself to Stripe at some point. Whether the rest of the zone is in good order is a separate question, and the same scan answers it for your own domain.

Figures as of 5 September 2026, from the September 2026 edition of the defaults.exposed census. Census numbers move every month; the current values are on the census data page.