What Are MTA-STS and TLS-RPT, and Does Your Domain Need Them?
Published 2026-11-14 · Reviewed 2026-11-14 · 6 min read · By the Ingenious AI team
Is someone impersonating your domain right now?
Run the free DomainGuard scan: lookalike registrations, MX-armed variants, certificate-transparency and CAA checks for your domain, in one report.
Run my free DomainGuard scanThe short answer
MTA-STS is a policy your domain publishes telling sending mail servers they must deliver your mail over verified encryption — and must not deliver at all if a proper TLS connection cannot be established. TLS-RPT is its reporting companion: aggregate reports from senders about encryption failures they hit delivering to you. Together they close the downgrade gap that SPF, DKIM, and DMARC do not cover — those three authenticate who mail is from; MTA-STS protects how mail to you travels. Deploy it if your mail provider supports it, after your certificates and MX are stable, in testing mode first.
The gap these standards fill
SMTP was built to deliver opportunistically: if encryption fails, traditional mail falls back to plaintext rather than failing. That fallback can be forced — an attacker positioned on the network path can interfere with the encryption negotiation (a downgrade or STARTTLS-stripping attack) and then read or alter mail in transit, including password resets and invoices, between servers whose owners both believe they run encrypted services. Authentication records never see this attack: the mail is genuinely from the sender; it was simply read on the way.
MTA-STS borrows the HSTS idea from the web: the domain declares, in advance and over authenticated channels, that its mail must arrive over TLS with a valid certificate for its MX hosts — and that senders should refuse delivery rather than fall back when that cannot be achieved. TLS-RPT gives you the telemetry to deploy that declaration safely, because senders report the failures they encounter, letting you fix your side before enforcement turns a misconfiguration into lost mail.
What deployment actually involves
Three pieces. First, every MX host must present a valid, unexpired, correctly-named TLS certificate — most reputable mail providers handle this, and it is the step to verify before anything else. Second, the policy file itself, served over HTTPS at a fixed address (mta-sts.yourdomain.com, at /.well-known/mta-sts.txt), listing your MX patterns, the policy mode, and a cache lifetime. Third, two DNS records: a TXT record at _mta-sts announcing the policy's version, and a TXT record at _smtp._tls carrying your TLS-RPT reporting address.
Modes mirror DMARC's philosophy. Testing mode asks senders to report but still deliver on failure — the only sane starting point, held for at least a few weeks of reports. Enforce mode instructs refusal on failure. The policy's max_age controls how long senders cache it, so start short (a day) while testing and raise it (weeks) once enforcing, remembering that a cached policy outlives your edits — plan MX changes around it exactly as you would a certificate pinning decision, because that is what it is.
Reading TLS-RPT reports and fixing what they show
Reports arrive as JSON, daily, from major senders, summarising successful and failed TLS sessions per MX host, with failure categories: certificate mismatch or expiry, negotiation failures, and policy-level issues. In testing mode your job is forensic comfort: every failure should be explainable — a legacy sender without TLS support, a misconfigured secondary MX, a certificate on one host renewing late. Unexplained failure classes are the deployment blockers to fix before enforce mode, because enforce converts them into bounced mail.
Expect the long tail honestly. Some legitimate senders, especially older on-premise systems at small partners, still cannot do modern TLS; enforce mode will refuse their mail. The reports tell you who they are while you are still in testing, which is the information you need for a grown-up decision: contact the partners, accept the risk window, or — for domains whose mail is mostly consumer-provider traffic, where TLS support is universal — enforce without ceremony. Most small companies on major mail platforms find the tail is empty.
How it composes with DMARC, SPF, and DKIM
The records protect different directions and different properties, and none substitutes for another. SPF, DKIM, and DMARC authenticate inbound claims about your domain — stopping others from forging you. MTA-STS and TLS-RPT protect delivery to your domain — stopping interception of genuine mail addressed to you. An estate with only the first trio can still have its inbound invoices read in transit; an estate with only MTA-STS still lets anyone spoof it. The complete email posture publishes all five, plus CAA on the certificate side.
Sequencing reuses the same muscle: inventory, publish in observe mode, read reports, fix, enforce. If you have run a DMARC rollout, an MTA-STS rollout will feel like a smaller echo of it — same discipline, fewer senders to chase, since the counterparties are receiving-side systems rather than your own marketing stack. DomainGuard's posture reporting tracks these records alongside DMARC so the whole email-authentication picture for a domain sits in one view.
Who needs it, and who can wait
Prioritise MTA-STS if your inbound mail carries the crown jewels: law, finance, health, property, and any business whose password resets and payment instructions arrive by email — which, candidly, is most businesses. The cost is low on a major mail platform (often an afternoon, mostly verification) and the reports alone justify testing mode, because undiagnosed TLS failures are otherwise invisible to you while being routine knowledge to anyone watching your mail path.
Wait, briefly, if your mail infrastructure is mid-migration or your MX certificates are in any way irregular — enforce mode on a shaky MX estate converts this security upgrade into an availability incident. Publish in testing, let the reports certify the estate, then enforce. As with every standard in this series, the staged path is not the slow path; it is the one that arrives.
Five actions to take today
- Verify every MX host presents a valid TLS certificate with a matching name before publishing anything.
- Publish the MTA-STS policy file over HTTPS and the _mta-sts TXT record, starting in testing mode with a short max_age.
- Publish the _smtp._tls record with a TLS-RPT address you will actually review.
- Read TLS-RPT reports for several weeks and explain or fix every failure class before considering enforce mode.
- Switch to enforce with a long max_age once reports are clean, and plan MX changes around the cached policy.
Worked example: Mesa Verde Property Group enforces MTA-STS
Mesa Verde Property Group, a fictional firm monitoring 8 domains, published MTA-STS in testing mode and read TLS-RPT for five weeks. Reports showed 99.1 percent successful TLS delivery; the failures traced to one legacy backup MX with an expired certificate — unknown to the team — and a small set of senders at two local firms using very old mail servers. The certificate was fixed in week two; the two firms were contacted and both, it turned out, had upgrades scheduled anyway. The firm switched to enforce with a 2-week max_age in week six, and its next quarter's reports showed zero unexplained failures across roughly 61,000 inbound messages.
FAQ
Does MTA-STS replace DMARC?
No — they protect different things. DMARC authenticates who mail claims to be from; MTA-STS protects the encryption of mail delivered to you. A complete posture publishes both.
What is the difference between MTA-STS and DANE?
Both enforce TLS for mail. DANE anchors trust in DNSSEC-signed records; MTA-STS anchors it in the web PKI and HTTPS. MTA-STS deploys without DNSSEC, which is why it is the practical choice for most domains today.
Will enforce mode block legitimate mail?
It refuses delivery where a compliant TLS session cannot be established — which is why testing mode and its reports come first. Fix what the reports show, and enforcement should be uneventful.
How long should we stay in testing mode?
Long enough to see every regular correspondent in the reports — typically three to six weeks, covering monthly senders. Seasonal correspondents argue for the longer end.
Our mail is with a major provider. Is this mostly done for us?
The provider handles certificates and TLS; your part is publishing the policy and DNS records and reading the reports. That is an afternoon of work and a standing review habit.