An HTML email is a web page written by someone else, and it opens next to your WordPress admin. A mail client in that spot has to decide what the page may do before a person reads it.
PressedMail does this in three layers. It cleans the HTML, it shows the result in a sealed frame, and it lets you decide when outside content loads. None of the layers can tell you whether the sender is honest, and this guide says where each one stops.
What happens to a message before you see it
First the server sanitizes the HTML. It keeps ordinary formatting such as tables, links and images. It removes scripts, forms, embedded frames and event handlers, limits links to safe schemes and drops the ping attribute that lets a link report clicks.
Then it neutralizes images that are at most two pixels wide and high, which is the classic tracking pixel. That catches the simple case. It does not catch a tracker hidden in a larger picture, and the tracking pixel guide explains why.
Finally the message goes into a sandboxed iframe. The only permissions the frame gets are to open links in a new window and to let that window leave the sandbox. It cannot run scripts, and it cannot reach WordPress cookies or the rest of the admin page. The server also sends a Content Security Policy with the document, and that policy does most of the day to day work:
- Images may only be data or inline attachments, until you allow remote images.
- Styles must be inline. Remote stylesheets are refused.
- Fonts must be embedded. Remote fonts are refused.
- The frame cannot load scripts, forms, objects, media or nested frames.
- The document is sent with a no-referrer policy, so a link you open does not tell the destination which admin page you came from.
A sandbox is explicit about what it allows. MDN’s iframe reference explains the attributes, and the DOMPurify project page describes the sanitizer the composer preview uses.
Who decides about remote images
Two switches control remote images, and they sit in different places.
An administrator sets the site rule. Open PressedMail, then Settings, then Access Control. The Email Display card has Allow remote images.

- Allow remote images, in the Email Display card. When it is off, remote images stay blocked across the inbox.
- Allow downloads, in the File Attachments card. Whether users can download attachments from received messages.
- Max attachment size. The largest file one attachment can be.
When Allow remote images is off, users still see a notice that a message has remote images. The reveal action is disabled, and each user’s own switch is hidden. When it is on, each user decides.
Each user decides under Settings, then Security, with Automatically load remote images. It is off by default. With it off, a message shows the notice and a Show images action, so you load pictures one message at a time.

- Your mailbox stays private. This is always on.
- Automatically load remote images.
- Store email on this site.
The note beside the switch says it plainly: loaded images may expose your activity to email senders through tracking pixels.
What happens when an image loads depends on the edition. In Free, a revealed image is requested straight from its HTTPS address by your browser. The paid builds send revealed images through a signed proxy on your own site, so the sender’s server sees your site’s request and not your browser. That is the paid plugin’s default, and there is no switch for it on the settings screens. Either way, the image’s host still receives a request. The remote image guide explains what that request can reveal.
Links and attachments
The guide to attachment limits and previews covers the download and size controls in detail.
Paid builds check each link locally while the message is rendered. They look for shortened URLs, raw IP addresses, lookalike domains, unusual ports and link text that names one place while the link goes to another. A warning page before a flagged link is available as a site option, and it is off by default. A flag is a reason to look closer, not a verdict. A flagged domain can be legitimate, and an unflagged one can be malicious.
Attachments are a separate decision. Viewing a message in a sealed frame says nothing about a file you download and open in another program. The File Attachments card in the screenshot above lets an administrator turn off uploads, Media Library attachments and downloads, and set the size limit. Check unexpected files before you open them.
What none of this proves
A message can pass every layer and still be a fraud. An invoice with changed bank details is ordinary text and an ordinary link. Rendering controls answer the question of what the page may do, and your own judgment answers the question of whether to believe it.
A phishing score, on the paid builds, is supporting evidence and not a clearance. The spam and phishing guide explains how to read one. For anything involving money or a login, confirm through a route you already trust.
A short check for site maintainers
If you maintain the site, test the real path after major updates instead of trusting a screenshot of a setting.
- Note the WordPress and PressedMail versions.
- Send yourself a harmless message with a table, an inline image and a remote image hosted on a server you control.
- Read it with automatic loading off, then on, and watch the browser’s network requests.
- Confirm that a link and a downloaded file stay separate actions.
- Review any custom code that renders message HTML.
This is a bounded check and not a security audit. Use your own test mail, never real correspondence, and remove addresses and tokens from anything you share when you report an issue.
The site around the inbox
The reading frame protects the page from the message. It does nothing about an out of date plugin elsewhere, a weak administrator account or an exposed backup. WordPress’s hardening guidance covers that work. On a shared site, decide who may open each mailbox, which the passphrase lock guide and the guide to where PressedMail keeps your mail cover. For client work, the agency mailbox guide shows how to keep accounts apart.
Frequently asked questions
Does sanitizing HTML stop all remote requests?
No. Sanitizing and the content policy are separate. Remote images load only after you allow them, and then the image host receives a request.
Does the iframe alone make email safe?
No. The permissions on the frame, the content policy, the sanitizer and the rest of the application together set the limits.
Can a safe-looking message still be phishing?
Yes. A fraudulent request can use plain text and a normal link.
Does Free include these protections?
Sanitizing, the sealed frame and remote image blocking are in every build. Link analysis, phishing reports and audit history are on the paid builds.
Manage your email inside WordPress
PressedMail Free adds an email client directly to your WordPress dashboard. Connect a compatible IMAP/SMTP mailbox, read and send mail, organize messages, and configure WordPress SMTP.


