The anti-spam that reads what a machine wrote.
More and more comment spam is written by a language model: it reads naturally, uses no spam vocabulary, and keyword filters cannot see it at all.
No account. No card. One click and your forms are protected.
Jessica Miller
jessica.seo@example.com
“Great article! It’s worth noting that in today’s digital landscape, the topics you delve into play a crucial role in shaping how businesses approach their online presence…”
- The writing style matches text generated by an AI model
- The form was filled in by automation, not by a person typing
Marta R.
marta@example.com
“Thanks — the bit about the cache TTL was exactly my problem. Fixed it and the page loads in under a second now.”
- Typing pattern is consistent with a real person
- This author has had comments approved here before
Numbers we actually measured
Every figure below was measured, not estimated. We publish the method and the misses too.
We will not tell you 99.9%. Nobody can measure that honestly, and the number that decides whether you keep a plugin installed is the first one — how often it blocks somebody real. Measured on 3,080 messages held out of training.
What you get
It sees the spam nothing else does
Machine-written comments have habits that human writing does not. We look for them directly, which is why we catch a class of spam that keyword and pattern filters are structurally blind to.
It will not block your customers
A blocked buyer costs far more than a spam comment sitting in a folder you already check. Everything is tuned around that, and suspected spam goes to the WordPress spam queue where a mistake is always recoverable.
It guards every door, not just comments
Comments, product reviews, registration, WooCommerce, Elementor, Contact Form 7, WPForms, Gravity, Fluent, Ninja, Forminator, Formidable and Mailchimp. Bots move to whichever form is unprotected.
Your database does not grow
No log tables. Nothing here turns into gigabytes that slow your backups and your admin. The plugin stores a score and a handful of reason codes per comment, and nothing else.
Analysed in Europe
Processing happens on European infrastructure. Comment text is not stored, and it is never sent to a third-party AI provider. What we keep is a one-way hash and a set of numbers.
Your visitors see nothing
No CAPTCHA, no puzzle, no "select all the traffic lights". If our service is slow or unreachable, local rules decide instantly and your forms keep working.
Activity
Checked and stopped, by month
Pricing
Free is a real product, not a trial. Most sites will never need more.
Free
€0forever
- 5,000 checks a month
- One site
- Every form integration
- Behavioural bot detection
- 30 days of history
Pro
★€9per month, per site
- 50,000 checks a month
- One site
- AI-written spam detection
- A year of history
- Priority support
Business
€49per month, per site
- 500,000 checks a month
- One site
- A model tuned to your traffic
- Webhooks and raw API access
- Two years of history
Upgrading takes effect on your next comment. The same key keeps working — there is nothing to reinstall.
Questions people actually ask
Will it block real comments?
About 0.9% of legitimate comments get held for review, measured on 3,080 samples. Suspected spam goes to the WordPress spam folder rather than being destroyed, so a mistake costs you a click. Commenters you have approved before are trusted automatically.
Does it work in my language?
Detection is strongest in English and Spanish. Other languages get the behavioural and structural layers, which are language-independent and catch the bulk of automated spam — but not well-written spam in a language we do not yet cover. We would rather say so than let you find out.
What happens if your service goes down?
Your forms keep accepting submissions. Local rules take over instantly, detection is weaker while that lasts, and nothing breaks. An anti-spam plugin that takes your comment form down when its backend hiccups is worse than no plugin.
Can I run it alongside Akismet or Wordfence?
With Wordfence, yes — different job entirely, they are complementary. With Akismet or another anti-spam, it works but you are paying twice in latency for overlapping checks. We never override another plugin’s verdict, so installing us can only add protection, never weaken what you have.
Is my visitors’ data sent to OpenAI?
No. Detection runs on our own infrastructure in Europe. No comment content reaches any third-party AI provider.
Does it need JavaScript?
The behavioural check does. A visitor with JavaScript off is never penalised for it — the other layers simply decide on their own.
I run several sites. What do I pay?
One subscription per site. Each install registers itself, gets its own key and its own monthly allowance, and is billed on its own — so five sites on Pro is five subscriptions. The free plan is per site too, and there is no limit on how many free sites you run.
Try it on one site
Install, click once, done. If it is not better than what you have, remove it — it leaves nothing behind.
Add to WordPress — free