Every AI-made ad, page and email is checked before customers see it

New

• Content check for AI agents. It stops copy that:
  • names a prescription drug, claims to cure or treat a disease, or names a medical condition
  • talks about prescriptions, doctors, pharmacies or online medical visits for a product that isn't a prescription medicine
  • uses another brand's name, or a page that loads pictures or links from another brand's site
  • shows a price, discount or deal your store doesn't sell, a review count or star rating your product doesn't have, or a number of customers nobody counted
  • promises shipping times, a money-back guarantee or subscription terms your policies don't give
  • shows placeholder text, a link that goes nowhere, or someone's personal email or phone number
  • tells a customer their order shipped when nothing shows it did
• Your own brands' content also gets a closer read: every claim about what the product does or contains must match the product's approved claims or label.

Worth knowing

• Write your product's approved claims, ingredients and words it must never use in the product's custom fields (group compliance). Claims that aren't there are stopped, never guessed.
• If the check can't finish, the content waits. It is never sent unchecked.

For developers

• POST /api/v1/agent/v1/verify and the MCP tool dtc_verify_content: send kind, texts (with role) or html, and product_url or product_id. The answer is verdict (pass, block or wait), layers and reasons (code, match, a short plain label and a detail sentence). Any key for the store can call it.