Named OpenAI, alongside Anthropic, as a provider of our AI text features in Section 9's list of companies that process data for us. Data we send to OpenAI's API is not used to train its models.
Said in Section 9 that Cloudflare also resizes and frames images for us. When a site's author chooses to frame a photo on people's faces, Cloudflare finds where the faces are so they stay in view. It does not identify anyone, and we never receive or store anything about the faces it finds.
Added Section 3.3, Website interaction information. If you send us a form, the pages you viewed during that same visit can be linked to your message so we understand what you are asking about before we reply.
Spelled out the limits on that link: it happens only when you send a form, it covers only that one visit, nothing is stored on your device, and only a random code reaches our analytics. Your name, email address, and message are never sent there.
Added a line to Section 4 so that analytics working without cookies is not read as analytics recording nothing.
Noted in Section 16 that customers can switch the same feature on for the sites they build, and that it is off unless they do.
Added Section 7, AI Features and Automated Processing. It describes what our AI tools do with your information, how the visitor AI concierge works, and our commitment never to use your content to train AI models.
Named Anthropic and Google as the providers behind our AI features, and added them to the list of companies that process data for us.
Explained that visitor conversations with the AI concierge are not stored, and that the concierge searches a text index we build ourselves rather than sending your content to a third-party search provider.
Added AI prompts and generated content to the list of information we collect, and AI features to the list of reasons we use it.