DMARC record: what each tag does, and the policy that stops spoofing
A DMARC record says what you ask receivers to do with mail that puts your domain in the From line without proof that you sent it. Without one, receivers must not apply DMARC to your domain at all [1], and nothing you publish asks them to refuse a forged message.
The short answer
A DMARC record is a TXT record at _dmarc.example.com, such as v=DMARC1; p=reject; rua=mailto:dmarc@example.com. It asks receivers what to do with mail that shows your domain in the From line and has no aligned SPF or DKIM pass: nothing (p=none), treat it as suspicious (p=quarantine) or refuse it (p=reject) [1].
Anatomy of a DMARC record
A DMARC record is a TXT record at the name _dmarc under your domain:
_dmarc.example.com. TXT "v=DMARC1; p=reject; rua=mailto:dmarc@example.com"
It is a list of tags and values separated by semicolons [1]. Only v is required. The policy in p does the protecting, and the rest refine it.
| Tag | Example | What it does | If you leave it out |
|---|---|---|---|
v |
v=DMARC1 |
Marks the record as DMARC. It must be the first tag. If it is missing, misspelled or not first, receivers ignore the whole record [1]. | Required |
p |
p=reject |
The policy for the domain, and for its subdomains unless sp or np says otherwise: none, quarantine or reject [1]. |
Treated as p=none if rua is set; without rua, as no DMARC at all [1] |
sp |
sp=quarantine |
The policy for subdomains that exist [1]. | p applies |
np |
np=reject |
The policy for subdomains that do not exist. RFC 9989 took it from the experimental RFC 9091 [1]. Receivers still on RFC 7489 ignore it as an unknown tag and apply sp or p [3]. |
sp applies, then p |
rua |
rua=mailto:dmarc@example.com |
Where receivers send aggregate reports [1][2]. | No reports at all |
ruf |
ruf=mailto:dmarc-failures@example.com |
Where receivers send per-message failure reports. Receivers concerned about privacy redact them heavily or send none [1]. | No failure reports |
adkim |
adkim=s |
DKIM alignment. r (relaxed) accepts a signing domain that shares your organizational domain; s (strict) needs the exact From domain [1]. A service that signs as mail.example.com passes relaxed and fails strict, as the mistakes to avoid explain. |
r |
aspf |
aspf=s |
The same choice for SPF [1]. | r |
t |
t=y |
Testing mode. Receivers apply one level below your policy: reject acts as quarantine, and quarantine as none [1]. |
n |
fo |
fo=1 |
Options for failure reports. Ignored without ruf [1]. |
0 |
psd |
psd=n |
psd=y marks the record of a public suffix, such as a top-level domain. psd=n marks a domain as its own organizational domain. Most domains leave it out [1]. |
u |
RFC 9989 removed three tags that RFC 7489 had: pct, rf and ri [1]. Receivers built on RFC 7489 still read pct [3]. From p=none to p=reject explains what that means while you tighten your policy.
Your DNS panel may split a long record into several quoted strings. Receivers join the strings before they read the record, so it still counts as one record [1].
How receivers find your DMARC record
A receiver takes the domain in the From line and looks for a TXT record at _dmarc in front of it [1]. For mail from news.example.com, that is _dmarc.news.example.com.
If nothing is there, the receiver walks up the DNS tree one label at a time and uses the record of the organizational domain, here example.com [1]. For the subdomain, it takes the sp policy if the subdomain exists and np if it does not. Without those tags, p applies [1].
Two DMARC records at the same name cancel each other. Receivers that follow RFC 9989 discard both and carry on up the tree as if you had none [1]. Receivers still on RFC 7489 stop there and apply no DMARC at all [3].
RFC 9989 replaced the Public Suffix List with this walk up the tree. Receivers still on RFC 7489 use the list and can reach a different answer for unusual names. RFC 9989 says strict alignment plus an explicit record on every domain you send from avoids the difference [1].
p=none, quarantine or reject: what each one does
| Policy | What receivers do with mail that fails DMARC | What the checker says |
|---|---|---|
none |
Nothing changes. RFC 9989 says a policy of p=none must not modify how mail is handled [1]. |
“Anyone can send email as example.com.” |
quarantine |
Treat it as suspicious. It does not reach the inbox at receivers that honor the policy [1]. | “example.com is partly protected.” |
reject |
Refuse it [1]. | “example.com is protected from forged email at receivers that check DMARC.” |
Receivers keep the final say. RFC 9989 lets them accept mail that fails even under p=reject, and a receiver with no other evidence must treat failing mail under p=reject as quarantine [1]. p=reject is still the strongest request you can make.
The checker counts some records one level lower than their p. t=y or pct=0 lower the policy by one level [1][3]. p=reject with sp=none or np=none leaves subdomains open, so the result is partly protected. And an SPF record that allows the whole internet, such as +all, makes the domain spoofable whatever its DMARC policy: all matches every address [4], so anyone can pass SPF with your domain as the envelope sender, and that pass aligns [1]. The About page lists every rule the checker applies.
Do you need DMARC if you already have SPF and DKIM?
SPF checks the envelope sender, an address the reader never sees [1]. A DKIM signature can come from any domain, including one an attacker owns [1]. Neither one ties the message to the From address the reader trusts, which is the address spoofing fakes.
DMARC adds that link, called alignment: the domain that passed SPF or DKIM has to match the From domain, or share its organizational domain under relaxed alignment [1]. The DKIM guide covers keys and selectors.
- A forged message arrives
- SPF
- DKIM
- DMARC
- What the policy asks:
p=noneor no recordNo requestp=quarantineTreat as suspiciousp=rejectRefuse
Large mailbox providers also ask for it. Gmail requires a DMARC record from bulk senders, those that send close to 5,000 messages or more to personal Gmail accounts in 24 hours [7], and accepts p=none for that rule [5]. Google’s sender guidelines word that line as more than 5,000 messages a day [5]. Without a record, Gmail rate-limits or blocks their mail with a message that starts like this one [6]:
Sender requirements compares what Gmail, Yahoo and Outlook.com ask for.
Problems and fixes
Your DMARC policy is p=none
The checker says “Anyone can send email as example.com.” although your domain publishes DMARC. Read From p=none to p=reject without blocking your own mail.
No reports, or zipped XML you can’t read
Your record has no rua tag, or your inbox fills with .xml.gz attachments from receivers. Read How to read a DMARC aggregate report.
A domain that sends no email
Without a DMARC record, receivers apply no DMARC to mail that uses the domain [1]. A parked domain, or one that only serves a website, needs p=reject and three more records. Lock down a domain that sends no email.
Check your DMARC record
The checker finds your DMARC record the way a receiver does, reads it tag by tag and tells you what it asks receivers to do with forged mail.
Runs in your browser. Your domain is not sent to this site or stored.
Guides in this topic
From DMARC p=none to p=reject without blocking your own mail
Move DMARC from p=none to quarantine and then reject without losing real mail. What to fix first, how long to wait, and how t=y, pct and subdomains behave.
How to read a DMARC aggregate report, line by line
A DMARC aggregate report is a zipped XML file from each receiver. Read its fields line by line, sort each row into your mail, forwarding or spoofing, and act.
