ETPSitesPreview, the reader behind ETP Sites' previews

ETPSitesPreview is the program that reads a school's website when someone asks ETP Sites for a preview of a new website for that school. It reads the school's current public website once, so we can show the school its own name, words and photos in our designs.

What it reads

• Only public pages. It never signs in, never fills in or submits a form, never runs a page's scripts, and never tries to get past a security check.
• Only when a person asks for a preview of that website. It does not crawl the web, follow links to other websites, or come back on a schedule.
• At most the site's robots.txt, its home page and up to 20 of its pages, 3 stylesheets and 13 images, within about 25 seconds.

How fast

• At most 4 requests at a time to one website.
• If the website's robots.txt sets a Crawl-delay, it waits that long between requests, and reads fewer pages rather than hurry.

robots.txt

ETPSitesPreview reads robots.txt before anything else and obeys it. It follows the rules a robots.txt gives for ETPSitesPreview, and otherwise the rules it gives for all crawlers (User-agent: *). To turn it away from your whole site, add these two lines to your robots.txt:

• User-agent: ETPSitesPreview
• Disallow: /

How it identifies itself

Every request carries the user agent ETPSitesPreview/1.0 (+https://etpcampus.com/etpsitespreview).

Every request is also signed with Web Bot Auth (HTTP Message Signatures, RFC 9421). Our public keys are published at preview.etpsites.com/.well-known/http-message-signatures-directory. A request that names ETPSitesPreview but is not signed with one of those keys did not come from us.

What happens to what it reads

• The preview has a random address, is not linked from anywhere, and is not shown to search engines. It is deleted after 14 days, unless someone saves it to their ETP Campus account: then it is deleted 90 days after anyone last opened it.
• The pages it reads are sent to an AI model (Anthropic's Claude) to write the preview. We do not use them to train any model or to build a search index.

For a school's IT

• To let it through a security service that challenges automated visitors, add a rule there that lets requests through whose user agent contains ETPSitesPreview. Your provider's documentation has the exact steps. Where your service checks Web Bot Auth signatures, it can also confirm that a request is signed with our keys.
• To turn it away, use robots.txt (above), or block that user agent in your firewall.

Contact

Questions about ETPSitesPreview, or a request from it you did not expect: support@etpcampus.com. ETPSitesPreview is run by ETP, Education & Technology Professionals.