About ShippingtonBot

Shippington builds a website assistant for a business from that business's own public pages. When someone asks us to build one, we read a handful of pages so the assistant can answer using what the site already says, rather than making things up.

Every request we send carries the header X-Shippington-Bot: https://shippington.co/bot, which is the link that brought you here.

It is not a crawler. We do not index the web, we do not follow links off your domain, and we do not come back on a schedule. A visit means a specific person typed your address into our site and waited for the result.

One visit reads at most six pages, once, from your domain only. Requests come from Google Cloud in us-central1, or from Anthropic's fetcher when a page needs JavaScript to render.

We used to name ourselves in the User-Agent string, which is the polite convention. It does not survive contact with commercial firewalls: several returned a flat 403 to a request that a browser was served normally, so a business asking us to read their own website got told we could not read it. We now send a browser user agent and identify in a header instead. That is a deliberate trade, and it is why the header exists.

We read and honour robots.txt before fetching anything. To exclude us specifically:

User-agent: ShippingtonBot
Disallow: /

A blanket User-agent: * rule applies to us too. If you would rather we never touch your domain again, or you want an assistant that was built from your site taken down, email vince@shippington.co and it is removed the same day, no questions and no argument.

• We do not log in, submit forms, or attempt to reach anything behind authentication.
• We do not read pages your robots.txt disallows.
• We do not follow links to other domains.
• We do not resell, republish, or train on your content. It grounds one assistant for the business that asked, and nothing else.
• We do not run on a schedule. No visit happens without a person asking for it.