Content-based spam filtering
An inbox filter that inspects an email's words, links, HTML, and layout for spam patterns, then combines those clues with other delivery signals.
See it
What it is
Content-based spam filtering inspects what is inside a message: its subject and body wording, HTML structure, links, images, attachments, and signs of obfuscation. A filter combines those clues into a spam decision or score. Apache SpamAssassin is the concrete, inspectable example: its rules add and subtract points, and mail above the configured threshold is marked as spam. Large mailbox providers use their own unpublished systems rather than one shared checklist.
Reach for a content audit when inbox placement drops after a copy, template, or link-tracking change. Compare the exact MIME message before and after the change, including the plain-text part, rendered HTML, redirect chain, and attachment types. That catches problems a quick read of the visible copy misses, such as a broken HTML part, a suspicious tracking domain, an image-only message, or a link whose displayed text does not match its destination.
There is no universal banned list of 'spam words'. One use of 'free' or an extra exclamation mark does not decide placement by itself, and polishing the copy cannot repair poor sender reputation, complaints, or failed authentication. A tool like mail-tester.com will run SpamAssassin over a message you send it and list the rules it tripped, but a clean score there is a useful diagnostic, not a promise that Gmail or Outlook will put the message in the inbox.
Ask AI for it
Audit the raw RFC 5322 email below for content-based spam risk. Run or interpret Apache SpamAssassin 4.0 with 'spamassassin -t', list every triggered rule and its score, and separate message-content findings from sender reputation and SPF, DKIM, or DMARC findings. Inspect the Subject, plain-text and HTML MIME parts, image-to-text balance, visible link text, final link destinations, tracking domains, redirects, attachments, and HTML validity. Then rewrite only the risky copy or markup while preserving the offer, required legal text, unsubscribe link, and tracking parameters. Return the revised plain-text part, revised HTML part, and a before-and-after table of SpamAssassin rules. Do not claim the new score guarantees Gmail or Outlook inbox placement.