To read email headers, open the message's full source, then work through the Received: lines from the bottom up and check the Authentication-Results: line for SPF, DKIM and DMARC. The From line of an email is just text that anyone can forge. The headers are different: they are written by the servers that handled the message, in order, and they record where it entered the mail system and whether it passed the checks that prove a domain really sent it.
Reading them takes a few minutes to learn and is the most reliable way to judge a suspicious email.
How do you open the full headers?
- Gmail: open the message, click the three-dot menu next to Reply, choose "Show original". Gmail also shows a summary table at the top with SPF, DKIM and DMARC results.
- Outlook (desktop): open the message in its own window, then File, Properties; the headers are in the "Internet headers" box. In Outlook on the web: three dots, View, "View message source".
- Apple Mail: View, Message, "All Headers", or "Raw Source" for the whole thing.
- Thunderbird: View, Message Source, or press Ctrl+U.
- iPhone and Android apps generally do not expose headers; forward the message to a desktop client.
You will see a block of lines in the form Name: value, often several dozen, newest at the top. Copy the whole block. Read it by hand with the guide below, or paste it into an email header analyzer, which lays the hops out in order with timing and highlights the authentication results.
The headers that matter
Received: the delivery chain
Each server that handles the message adds a Received: line to the top of the headers. Reading from the bottom up therefore follows the message from its origin to your inbox. A typical line looks like this:
Received: from mail-sor-f41.google.com (mail-sor-f41.google.com. [209.85.220.41])
by mx.example.net with ESMTPS id k7si2290871
for <you@example.net>; Tue, 7 Oct 2026 09:14:22 -0700 (PDT)
"from" names the server that handed the message over, with the IP address it connected from in brackets; "by" names the server that received it; the timestamp says when. The bottom-most Received line shows where the message entered the system. If an email claims to be from your bank but the first hop is a residential broadband address, you have your answer.
Two cautions. The sender can add fake Received lines below the real ones, because they control the message before it is sent; only lines added by servers you trust, starting with your own provider's, are guaranteed genuine. And large providers route mail through several internal servers, so five or six hops from Google or Microsoft is normal.
Return-Path: where bounces go
This is the envelope sender, the address given in the SMTP conversation, which is often different from the From header. Bulk senders use a unique Return-Path per recipient to track bounces. SPF is checked against this domain, not the From domain.
From, Reply-To and Sender
From: is what you see in your client. Reply-To:, if present, is where replies actually go; a Reply-To that differs from the From domain is a classic sign of a scam. Sender: appears when a message is sent on behalf of someone, for example by a mailing-list service.
Message-ID, Date and X- headers
Message-ID is a unique identifier assigned when the message was created, and its domain part often reveals the real sending system even when the From address is forged. Date is set by the sender's software and can be wrong or faked; compare it with the Received timestamps. Headers beginning with X- are non-standard extras: X-Mailer names the software used and X-Originating-IP sometimes records the author's IP.
A real header block, line by line
Here is a message we actually sent while updating this article. It is an internal test: a script on our own server delivered it to a fresh MailboxTemp inbox over SMTP, labelled as such in its subject. These are the headers exactly as the sending script wrote them:
From: MailboxTemp test <verify@mailboxtemp.com>
To: <wrl4u9a@fakhita.xyz>
Subject: [internal test] Your verification code
Message-ID: <test-1793549244512@mailboxtemp.com>
Date: Sun, 01 Nov 2026 16:07:24 GMT
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
- From carries a display name and an address. Only the part in angle brackets is an address; the name is free text, which is why "Your Bank <random@gmail.com>" is possible.
- To is also just text. It need not match the address the message was actually delivered to, which lives in the envelope and shows up in
Received … foror inDelivered-To. - Message-ID here is a millisecond timestamp plus the sending domain, which is typical: the right-hand side names the system that created the message, not the one in the From line.
- Date is the sender's clock, to the second. The receiving server recorded the same message at
16:07:24.639Z, so sender and receiver agreed within a second. A Date hours away from the first trusted Received timestamp means a wrong clock or a backdated message. - MIME-Version and Content-Type tell the client how to decode the body: plain text, UTF-8, no HTML part and no attachments.
Notice what is missing. There is no Received line, no Return-Path and no Authentication-Results, because the sender never writes those; the servers that handle the message do. Our receiving server, Haraka, stamps one Received line as the message arrives. Its template, from Haraka's own source, filled with the values from this delivery's SMTP log, reads:
Received: from mail.mailboxtemp.com (ip6-localhost [127.0.0.1])
by mail.mailboxtemp.com (Haraka/3.2.1) with ESMTP
id 242A6D5B-415C-496C-8ADC-3F2F6EEAC9A2.1
envelope-from <verify@mailboxtemp.com>; Sun, 01 Nov 2026 17:07:24 +0100
The name after from is whatever the sender announced in its HELO command; the bracketed address is the connection the server actually saw, and that is the part to trust. Here the two disagree, a server name against a loopback address, exactly the pattern a forged hop would show. with ESMTP rather than ESMTPS means this hop was not encrypted, and envelope-from is the value that becomes Return-Path. The timestamp is in the server's local zone, +0100, while the Date header above was written in GMT: the same second, two notations, which is why you compare offsets before calling a clock wrong. There is no Authentication-Results line at all: an unsigned message from the loopback interface has nothing for SPF or DKIM to evaluate, so an honest server writes nothing rather than a hollow pass. A message from Gmail arriving at the same server would carry one.
Two practical notes. The inbox's public API returns the parsed fields, from_address, from_name, subject, otp_code and the server's received_at, rather than the raw header block, so for Received-chain forensics use "Show original" in a full mail client. And a seven-line test header is readable by hand; a real message with eight Received lines, three Authentication-Results entries and a dozen X- headers is where pasting the block into the header analyzer stops being a convenience and starts preventing mistakes.
What do the SPF, DKIM and DMARC results mean?
Modern receiving servers check three things and write the results into an Authentication-Results: header. This is the most useful line in the whole block.
Authentication-Results: mx.example.net;
spf=pass smtp.mailfrom=bounces.shop.com;
dkim=pass header.d=shop.com;
dmarc=pass header.from=shop.com
- SPF (Sender Policy Framework) asks: is the IP address that delivered this message on the list the Return-Path domain publishes in DNS?
passmeans yes.failmeans an unauthorised server sent it.softfailandneutralmean the domain's policy is lenient; treat them as "unknown". - DKIM (DomainKeys Identified Mail) asks: was this message cryptographically signed by the domain in
header.d, and is it unmodified since? Apassproves the message really came through that domain's mail system. Afailusually means the message was altered in transit, often by a mailing list, or forged. - DMARC asks: does the domain in the visible From header match the domain that passed SPF or DKIM, and what does the From domain's policy say to do if not?
dmarc=passis the strongest signal that the From address is genuine. A domain whose DMARC policy isp=rejectcannot be easily impersonated; one with no DMARC record can.
The crucial subtlety: these checks verify that the message came from the domain it says it came from. They do not verify that the domain is trustworthy. A phishing email sent from secure-paypal-login.com will pass all three with flying colours, because the attacker set them up correctly. Authentication tells you the sender is who they claim; you still have to decide whether that claimant is the real company. Our guides to SPF, DKIM and DMARC and to spotting phishing emails cover the rest of that judgement.
A worked example
Suppose a message arrives apparently from billing@yourbank.com saying a payment failed. In the headers:
Authentication-Resultsshowsspf=pass smtp.mailfrom=mail.promo-blast.net,dkim=pass header.d=promo-blast.net,dmarc=fail header.from=yourbank.com. The message was genuinely sent by promo-blast.net, but that domain is not authorised to use yourbank.com in the From line. DMARC failed because the domains do not align.- The bottom
Receivedline shows the message entered the mail system from an IP address belonging to a cheap hosting provider, not the bank. Reply-Tois a webmail address.
Each of those alone is suspicious; together they are conclusive.
Reading headers from your own mail
Headers are also the fastest way to debug mail you send. If your newsletters land in spam, the Authentication-Results on a test message tells you whether SPF, DKIM and DMARC are set up correctly. If delivery is slow, the gaps between Received timestamps show which hop is holding the message. Developers testing signup emails can send to a disposable test inbox for content and timing, then check authentication at a mainstream mailbox that exposes the raw source.
Limits of header analysis
Headers show the path a message took and whether it was authenticated. They do not reveal the person behind a free webmail account, since the originating IP is the provider's, and they cannot tell you whether a correctly authenticated domain is run by honest people. They are one input to a judgement, not a verdict.
In short
Email headers are the server-written record of where a message came from and whether its domain was verified. Open the source, read the Received lines from the bottom up, trust the bracketed IP over the announced name, compare Return-Path, From and Reply-To, and trust the Authentication-Results line added by your own provider. A DMARC failure on a message that asks you to act is your cue to verify through another channel.
Sources
- RFC 5322, Internet Message Format, which defines the From, Reply-To, Sender, Date, Message-ID and Received header fields.
- RFC 5321, Simple Mail Transfer Protocol, section 4.4, the trace information a receiving server must add in its Received line.
- RFC 8601, Message Header Field for Indicating Message Authentication Status, the specification of the Authentication-Results header.
- RFC 7489, Domain-based Message Authentication, Reporting, and Conformance (DMARC).
- Haraka, the SMTP server whose Received-line format is quoted above.
- Gmail Help: Trace an email with its full header.
- Microsoft Support: View internet message headers in Outlook.