If LetThemBuild appears in your logs, a person using LetThemBuild or LetThemChat asked the AI seats in their discussion to read one of your pages, and the room fetched it for them. This page says what that is, what it is not, and how to refuse it.
What it fetches, and why
- One page at a time, the one a person named or a search result they asked the room to read. Each fetch is a person's request, made once, on their behalf.
- It reads the page as text so the models in the room can answer from it. What the page says is treated as data the seats read, never as instructions they follow.
- It never crawls. It does not follow your links, walk a sitemap, index your site, or come back on a schedule. There is no store of your pages on our side: the text lives in that person's discussion and nowhere else.
- Nothing fetched here trains a model.
How it identifies itself
Every fetch carries this user agent, on the person's Mac and on our servers alike:
LetThemBuild/1.0 (+https://letthembuild.com/bot)
How to refuse it
The bot reads your robots.txt before every page and honours it: a path you close to LetThemBuild, or to every agent, is not fetched, and the person is told that your site does not allow automated reading of it. To refuse it everywhere:
User-agent: LetThemBuild Disallow: /
Or close one part of the site and leave the rest open:
User-agent: LetThemBuild Disallow: /members/ Allow: /
A robots.txt the bot cannot read is taken as permission, because a reader acting for one person is not a crawler; the moment your file answers, its rules apply. Changes are picked up within an hour.
Where to write
If the bot did something this page does not describe, or you want a site left alone without editing robots.txt, write to [email protected] with the host and the time from your logs. We answer from the same address.