AI Anti-Spam & Bot Detection: bots stopped before the form
Most anti-spam plugins decide after the message is typed. This one works earlier: it looks at the visitor, the browser and the request, and blocks the ones that fail the check before a form is even rendered.
Three layers
Fingerprinting answers who is on the other side: browser, device, rendering quirks, a stable visitor ID that survives a cookie reset. The PHP library works on the server, the JavaScript library on the client, and both speak the same fields.
Behaviour answers how the visitor acts: typing speed, mouse path, time on page, the order of focus events. A script that fills a form in 400 ms is easy to separate from a person.
Content answers what the message says: pattern matching on links, repeated submissions and known spam signatures, with a model for the cases a rule misses.
What keeps it usable
A CAPTCHA punishes everyone. Here CAPTCHA appears as a fallback for uncertain traffic, not as the first line. Known good visitors pass without a click, which matters most on checkout and login forms.
The plugin also leaves the crawler question alone, and that is deliberate. Robots rules belong to a separate tool: WP Robots Generator writes robots.txt from the admin with security presets, plugin presets, AI crawler blocking and translations, without a backend or a dependency.
Measuring the result
Two numbers decide whether the setup works: spam that reached the database and real users who could not submit. Both are visible in logs, and both need to be checked in the first week.
Rules go in narrow and get wider. A rule that blocks every request from a cloud provider will also block a partner on a cloud VPN.
Where it fits
WooCommerce checkouts, comment sections, contact forms, registration and password reset. Each has a different tolerance: a store loses money on a false positive, a blog loses patience.
Related: BotBlocker Security for the firewall layer, WordPress security for the audit. Code: Fingerprint JS, Fingerprint PHP, WP Robots Generator.
