Per-page AI crawler controls

Pages now carry three AI opt-out flags, readable on the page payload and writable on create and update. block_ai_crawler blocks the crawlers that train models on a page (GPTBot, ClaudeBot, Google-Extended), block_ai_search blocks answer engines from indexing it and citing it (OAI-SearchBot, PerplexityBot), and block_ai_assistant blocks assistants from opening it when a person asks them to (ChatGPT-User, Claude-User).

Watch the polarity. seo_index is an allow flag, the three block_ai_* fields are block flags, and all four default to false, so a page starts out hidden from search engines but open to every AI category. The four are independent: seo_index does not gate any of them. On an external page all three are rejected with a 422, along with the other internal-only fields.

Blocks are expressed as Disallow rules in the domain's robots.txt, the surface AI crawlers actually honor (meta tags and the sitemap do not apply to them). Coverage follows the paths ClickFunnels can currently resolve for a page, which today means the funnel step paths of a page that is a step in a funnel. A page that is not a funnel step, such as a standalone landing page, site page, theme page, blog or course page, stores the preference and reads it back but contributes no rules yet. A 200 is therefore not proof that a page is blocked: fetch the domain's /robots.txt to confirm. Enforcement widens as path resolution is extended, with no further action on your side. There is no bulk endpoint, so opting a whole workspace out means listing its pages and PATCHing each one.

Agent examples: opt a launch funnel out of model training while leaving answer engines free to cite it · audit which pages in a workspace are still open to AI crawlers and flip the ones that should not be.

Resources: Page
Building with an AI agent? Point it at the Pages skill.