How the AI spam detector works
Traditional spam filters count suspicious words or block lists of addresses. Spammers get around both by changing a few letters or using a new link. This detector reads the content the way a person would: does this message make sense for the place it was posted, or is it an advert, a scam or a junk link in disguise?
What the model looks at
You paste a comment, contact form message, review, email or forum post and tell the tool what kind of content it is. The language model then looks at the meaning and purpose of the text: unrelated promotion, "guaranteed" income claims, pressure to click, generic praise that says nothing about the page, copied text and links that have no reason to be there.
It returns a verdict, a spam score and a short reason, so you can see why a message was flagged instead of getting a mysterious yes or no.
Content type and strictness
The same sentence can be spam in a comment and normal in a sales enquiry, so the content type matters. Strictness lets you choose the trade-off. Low flags only obvious spam and lets borderline messages through. High flags anything suspicious and may catch genuine messages. Medium suits most websites.
Toxicity and personal data checks
You can also ask it to flag toxic or abusive language and to note when a message contains personal information such as a phone number or email address. These are separate signals from spam, which is useful when you moderate public comments.
What it is not
This is a content checker. It does not test whether an email will land in the inbox, score a newsletter for spam trigger words or check a sending domain. For that you need a deliverability tool.
Using it on your own forms
The spam detection API has a free monthly quota on every plan, and RSFlowHub Forms already uses it to filter contact form messages. No detector is perfect, so keep a way to review flagged messages instead of deleting them silently.