Do Not Disturb My Comments AI reads every new comment and asks an AI model one simple question: is this spam?

If the answer is a confident yes, the comment is turned away before it ever reaches your database. You won’t find it in your spam folder because it was never saved. If the comment looks genuine, it’s approved straight away, so real readers aren’t left waiting. If you’d rather read every comment first, you can switch that off.

The checking is done by Jev, a decision model made by TypeSafe AI and offered through Cloudflare. Jev doesn’t write text. It returns a probability, which makes it fast; usually well under a second.

What it does:

Blocks comments Jev scores as likely spam, before they’re saved
Approves comments Jev considers genuine, so real readers aren’t left waiting (you can switch this off)
Skips the check for anyone who can moderate comments, like you and your editors
Leaves WordPress in charge whenever Jev can’t be reached, so nothing breaks
Respects your Disallowed Comment Keys. If WordPress already flagged a comment, it stays flagged

What it looks at when judging a comment:

Promotional messages and link dropping
Phishing and scams
Generic praise that has nothing to do with your post

What you need

Nothing, for most sites. The plugin comes with a free built-in spam-check service, so there are no accounts or keys to set up.

The built-in service checks up to 20 comments a minute for each site. That’s far more than most blogs ever get. Comments over that limit aren’t lost. They simply go through your normal WordPress moderation.

If you’d rather use your own Cloudflare account, you can. You’ll need a small prepaid balance on Cloudflare’s AI Gateway, because Jev is a third party model and Cloudflare’s free Workers AI allowance doesn’t cover it.

External services

This plugin sends comments to an outside service to decide whether they’re spam. Which service depends on the connection you choose on the settings page. Nothing is sent to either one until a site administrator saves the settings page.

After that, a request is made each time someone who can’t moderate comments submits a comment. Each request sends:

The title of the post being commented on
The commenter’s name
The commenter’s website address, if they gave one
The comment text

The commenter’s email address and IP address are not sent.

Built-in service (the default)

Comments are sent to the Do Not Disturb My Comments AI spam-check service at https://dndmc-proxy.chidideveloper.workers.dev/v1/check. It’s run by the plugin author on Cloudflare Workers. The service passes the comment to Jev, a model by TypeSafe AI, through Cloudflare, and sends back only a spam score. Comments are not stored by the service. The service sees your web server’s IP address, which it uses to apply the fair use limit.

Service terms: Terms of Service
Service privacy policy: Privacy Policy
Cloudflare terms: Cloudflare Terms
Cloudflare privacy policy: Cloudflare Privacy Policy
TypeSafe AI terms: TypeSafe AI Terms
TypeSafe AI privacy policy: TypeSafe AI Privacy Policy

Your own Cloudflare account

Comments are sent straight to Cloudflare’s Workers AI API at https://api.cloudflare.com/client/v4/accounts/{your-account-id}/ai/run, which runs them on Jev. Usage is billed to your Cloudflare account. The Cloudflare and TypeSafe AI links above apply.

Because comment content is shared with these services, consider mentioning it in your site’s privacy policy.