KlaviyoAIBot

KlaviyoAIBot gathers website content for Klaviyo's Kai Customer Agent. During setup, it can index a connected storefront, and a Klaviyo customer can add other sources such as a help center, blog, or individual webpage. Klaviyo's bot documentation describes the crawler and its controls.

The scope comes from sources connected to the Klaviyo account. Klaviyo says the bot fetches public pages from those domains and URLs rather than searching the open web for loosely related material. It does not bypass authentication, paywalls, or other access controls.

Indexed text becomes company knowledge for the customer agent. It can support answers about products and policies, content generation, and product recommendations. Klaviyo periodically revisits connected sources so that changes can replace older indexed material.

This is an account-bound retrieval workflow, not a documented foundation-model training crawler. A visit can affect what a merchant's Customer Agent says to shoppers, but it should not be presented as evidence that the page was added to a general training dataset or a public AI search index.

KlaviyoAIBot uses cryptographic identity instead of a published IP allowlist. Every request is signed with HTTP Message Signatures, and Cloudflare's bot directory lists the public key directory as https://www.klaviyo.com/.well-known/http-message-signatures-directory. An unsigned request claiming this name should be treated as untrusted.

Robots behavior remains unknown in this bot. Klaviyo's help page says the crawler follows the Robots Exclusion Protocol, while Cloudflare's directory entry marks it as not following robots.txt and supplies no user-agent pattern. That conflict makes source selection, signature-aware edge rules, and observed request behavior more reliable than an assumed token.