# Only the landing is public. Every other path is a personal, token-scoped page
# (/{token}, /{token}/api, /{token}/mcp) that must never be crawled or indexed —
# the token in the URL is the user's private key. Longest-match rules: the
# specific Allows below beat the blanket Disallow.

User-agent: *
Allow: /$
# ?lng= URLs are deprecated and now 308 to /<lang>/ (see vercel.json). This Allow deliberately
# STAYS until the duplicates leave the index: Google indexed /?lng=es, /?lng=fr and /?lng=ru and
# picked the parameter URL as its own canonical over the "/" we declared, so /?lng=ru was ranking
# at position 1.0 while /ru/ sat at 8.8. A robots block cannot fix that — a disallowed URL is never
# recrawled, so the redirect would never be seen and the entry would linger as a URL-only result.
# Remove this line only once URL Inspection reports all three as redirects rather than indexed.
Allow: /?lng=
Allow: /pt/
Allow: /ru/
Allow: /es/
Allow: /fr/
Allow: /guides/
Allow: /privacy/
# Support (the contact form): fetchable so a crawler sees its noindex and follows its links; not in
# the sitemap (see renderSupport in guides/template.mjs).
Allow: /support/
# Terms: same as support, noindex and not in the sitemap (see renderTerms).
Allow: /terms/
Allow: /index.html
Allow: /assets/
Allow: /icon.svg
Allow: /icon-maskable.svg
Allow: /apple-touch-icon.png
Allow: /pwa-192.png
Allow: /pwa-512.png
Allow: /pwa-512-maskable.png
Allow: /og-cover.png
# Guide covers, hero images and the promo film: the pages reference them, so crawlers must be able
# to fetch them for rendering, image search and link previews.
Allow: /og/
Allow: /guide-hero/
Allow: /promo/
# The exercise frames (/exercises/<slug>-N.webp). The exercise libraries (/fr/exercices/<slug>/ and
# /es/ejercicios/<slug>/, under the Allow: /fr/ and /es/ above) show them as their two images, so
# crawlers must be able to fetch them
# for rendering and image search. Only static images live here: the app's exercise screens are
# /{token}/exercises/..., which this does not open.
Allow: /exercises/
# The planilha guide's downloadable spreadsheet and PDF (/guides/pt/planilha-de-treino-excel-gratis/).
# The files are the page's answer to its query, so a crawler must be able to fetch them.
Allow: /downloads/
# The demo, and the connection films and step screenshots the connect guides embed. Every guide
# links to /demo/ since the signup-form rollout, and the connect guides reference the media, so a
# crawler that cannot fetch them sees a page it cannot render and a link it cannot follow. /demo/
# must stay fetchable but must not be indexed: it serves the landing shell with a canonical to "/",
# and Google ignored that canonical four days after this Allow went live (URL Inspection on
# 2026-09-26: indexed, Google-selected canonical /demo/). The noindex lives in vercel.json as an
# X-Robots-Tag header on "/demo/(.*)", not here - a URL blocked in robots.txt is never fetched, so a
# noindex on it would never be seen.
# Note the trailing slashes: the app's own /connect page stays disallowed, only its media opens.
Allow: /demo/
Allow: /connect/
Allow: /instructions/
Allow: /manifest.webmanifest
Allow: /llms.txt
Allow: /sitemap.xml
# Domain-ownership proof for the Glama connector listing. Must stay fetchable for as long as
# we want the listing to stay claimed - Glama re-checks it.
Allow: /.well-known/glama.json
# Bing Webmaster Tools site-ownership proof. Must stay fetchable while the property is
# verified; Bing re-checks it.
Allow: /BingSiteAuth.xml
# IndexNow key (web/scripts/indexnow.mjs). Bing and Yandex fetch it to check that a URL
# submission really comes from this site. Public by design.
Allow: /61732ef11730fee8318939243178b374.txt
Disallow: /

Sitemap: https://aim-journal.com/sitemap.xml