Knowing an email landed in spam does not tell you what to change. Content analysis reads the message that actually arrived, marks the exact lines working against you, and tells you which ones to fix first. Most of the time it is not your servers. It is a subject line.
Verdict. Authentication and delivery path are clean. The score is held back by wording and link balance, which is a content fix rather than a reputation one. Example report.
This is the report, not a description of it. Click any marker to see what the engine found, why it matters to a filter, and the specific change to make.
Hi there, act now, this offer expires today and you will not see it again.
To claim it, visit your accountgoes to bit.ly/2xKp9
Unsubscribe
All caps, an exclamation mark and "FREE" in the first three words. This is the single strongest spam signal in the message, and filters weight the subject heavily because it is the cheapest thing for a bad sender to abuse.
Say what the email is. "Your March invoice is ready" gets opened by the people who actually wanted it.
Example scores. The point is that none of these fixes touched your servers, your DNS or your reputation. They were all in the message.
Each one answers a different question, and you see all five results rather than a single number you have to trust.
We re-run the cryptographic checks on the message that arrived rather than trusting the headers it carried. If your signature broke somewhere in transit, that is a silent deliverability leak and you would never see it from your own sent folder.
Catches: broken DKIM signatures, SPF that fails from your real sending IP, DMARC that does not line up
We walk the delivery path backwards to find the machine that actually sent it and whether every hop was encrypted. Sometimes the answer explains everything: mail leaving from an IP you did not know you were using.
Catches: unexpected sending IPs, unencrypted hops, relays you did not intend
Around 105 rules over the subject, the wording, the HTML, the links and the images. This is the pass that produces most of what you can fix today, which is why it counts for the most.
Catches: shouty subjects, fake urgency, mismatched links, image-only emails, missing unsubscribe
We run the message through Rspamd, an open-source filter used in production by real mail servers, and show you every rule it triggered with the points attached. It is the closest thing to a second opinion from the other side.
Catches: the exact symbols a receiving filter would fire, and how close you are to its threshold
An AI pass reads the email the way a recipient would and says what a rules engine cannot: whether the tone is off, which Gmail tab it expects to land in, and what to write instead.
Catches: tone problems, readability, compliance risks, and specific replacement wording
The wording rules carry the most weight in the score, because they are the findings you can act on this afternoon without touching your infrastructure.
Links are where well-meaning email most often trips a filter. These checks run on every analysis.
Deeper link inspection exists but is currently switched off: following redirect chains, spotting dead links, unresolvable domains and blocklisted hops. It is held back because fetching arbitrary URLs from a report needs careful protection against server-side request forgery.
We would rather tell you a check is off than imply it ran.
There is no separate step. Run a placement test, and once a seed inbox receives the message the analysis starts on its own.
Send from your real sending path to the seed network.
The delivered copy is captured, exactly as a receiver got it.
Delivery path, authentication, rules, Rspamd and the AI review.
One composite out of 100, the full breakdown, a verdict and prioritized recommendations.
If nothing reaches a seed inbox, the report says so rather than guessing. That is a delivery problem, and blacklist monitoring plus Inbox Boost are where you fix it.
Run a placement test and the content analysis comes with it. Free to start, no card required.