# Let Agents In measures whether sites welcome agents. It would be absurd to block them here.
User-agent: *
Allow: /
Allow: /v/
Allow: /methodology
Allow: /corpus.json

# One scan each, superseded the moment we rescan, and there are thousands. The page for a
# company is /v/<domain> and it always holds the newest measurement. Meta's crawler read
# seventy of these in parallel on 2026-08-12 and took the site down, which is a fair thing to
# have learned about ourselves from the outside.
# Zostaje mimo `noindex` na tych stronach, i to jest swiadomy wybor: crawler, ktory posluchal tego
# Disallow, nigdy nie pobierze strony, wiec nie zobaczy tez `noindex` i moze zostawic w indeksie sam
# adres. Adres jest nieodgadywalny i sam z siebie nie mowi nic, a zdjecie tej linijki zaprasza z
# powrotem siedemdziesiat rownoleglych pobran, ktore raz juz polozyly ten dyno. `noindex` jest tu dla
# tych, ktorzy tego pliku nie czytaja (codex).
Disallow: /r/
# Personal links from an email. Nothing here is secret and none of it is worth an index entry.
Disallow: /watch/

Content-Signal: search=yes, ai-input=yes, ai-train=yes

Sitemap: https://letagentsin.com/sitemap.xml

# The agentic catalog, so a client can read what we expose before invoking any of it. ARD is barely
# adopted today; we publish it because we tell vendors to and it costs one static file.
AI-Catalog: https://letagentsin.com/.well-known/ai-catalog.json