Words are evidence,
not a spam verdict.

No public list of forbidden words can predict placement.

Filtering systems can evaluate authentication, IP and domain reputation, complaint history, recipient behavior, headers, URLs, message structure, content, and combinations of signals. A legitimate promotion can accurately say “free shipping.” A deceptive message can avoid every phrase in a static list.

This tool therefore identifies exact review evidence. Its index represents how much operator attention the draft needs under this disclosed ruleset—not the probability that Gmail, Outlook, Yahoo, an enterprise gateway, or SpamAssassin will classify it as spam.

Specific truth is a stronger rewrite than euphemism.

Do not merely swap a flagged phrase for a synonym. If the underlying claim, pressure, or identity is misleading, the message remains misleading.

Link language and link destination should agree.

HTML mode extracts anchors, image attributes, hidden elements, scripts, forms, iframes, shortener hosts, apparent tracking pixels, and visible-URL mismatches. A hidden element can be a legitimate preheader; a tracking pixel can be ordinary analytics. Both still belong in the evidence record.

Google's sender guidance says links should be visible and easy to understand so recipients know what to expect. Test the provider-compiled email because redirect and tracking systems can rewrite destinations after authoring.

A text scan cannot certify commercial-email compliance.

For marketing mode, the tool checks for visible opt-out language and a directional postal-address pattern. It cannot verify whether headers implement RFC 8058 one-click unsubscribe, whether an address is valid, whether an opt-out works, or which jurisdiction applies.

In the United States, the FTC's CAN-SPAM guidance requires accurate headers, non-deceptive subjects, ad identification, a valid physical postal address, a clear opt-out method, and timely honoring of opt-outs. Other countries and states can impose different or additional consent and privacy rules.

Fix the message system, not just the highlighted words.

1. Confirm recipient expectation.

   Document how and why this audience expects this message type.

2. Validate identity and claims.

   Make the sender recognizable and substantiate every material promise.

3. Inspect the compiled artifact.

   Review links, headers, authentication, HTML, plain text, and unsubscribe behavior after provider processing.

4. Send a controlled test.

   Use representative inboxes and preserve authentication and rendering evidence.

5. Monitor receiver feedback.

   Complaints, bounces, deferrals, and provider reputation are stronger operational signals than a word score.

Spam-language checker FAQ.