# sweat.care robots.txt
# Source of truth for the subdomain map: SUBDOMAINS.md
# Last updated: 2026-05-26
#
# This file is served identically from every subdomain on the sweatcare-admin
# Pages project. Admin subdomains rely on AuthGate to block unauthorized
# crawler access; the Disallow list below discourages crawling of known
# admin paths regardless of which host the crawler reaches.
#
# research.sweat.care is a separate Pages project (sweatcare-research) and
# manages its own robots.txt — total ban there, handled in that repo.

User-agent: *

# ─── Admin / internal paths — do not crawl ──────────────────────────────────

# Creator program admin
Disallow: /admin
Disallow: /all
Disallow: /applications
Disallow: /tracking
Disallow: /sourcing
Disallow: /partner
Disallow: /agencies

# Ops
Disallow: /orders
Disallow: /payouts
Disallow: /inventory
Disallow: /discounts
Disallow: /pos
Disallow: /finance

# Strategy / investor / fundraise (allowlist-gated)
Disallow: /build
Disallow: /deck
Disallow: /capital
Disallow: /channels
Disallow: /pipeline
Disallow: /pharma
Disallow: /revenue
Disallow: /road-to-la
Disallow: /brief/

# Media production / internal tooling
Disallow: /studio
Disallow: /voice
Disallow: /production
Disallow: /calendar
Disallow: /console
Disallow: /watch
Disallow: /wargame
Disallow: /experiments
Disallow: /market
Disallow: /customer
Disallow: /engagement
Disallow: /roadmap
Disallow: /research-agent
Disallow: /partnership-agent
Disallow: /sales-agent
Disallow: /mockups
Disallow: /insights-preview
Disallow: /insights-locked
Disallow: /social
Disallow: /brand

# Affiliates admin
Disallow: /affiliates

# Coach onboarding (auth-required flows)
Disallow: /welcome
Disallow: /feedback

# Personal / auth-required
Disallow: /account
Disallow: /r/
Disallow: /reveal
Disallow: /survey

# Test surfaces
Disallow: /shipping/test
Disallow: /checkout/test

# ─── AI policy: CITED yes, TRAINED no (AC, 2026-07-07) ──────────────────────
# Answer / citation crawlers (OAI-SearchBot, ChatGPT-User, Claude-User,
# Claude-SearchBot, PerplexityBot, Perplexity-User, Googlebot, Bingbot,
# Applebot) are NOT named here, so they inherit the User-agent:* rules above —
# allowed on public pages, blocked from admin. TRAINING crawlers are blocked
# outright below. (Edge note: also confirm Cloudflare's blanket "Block AI
# Scrapers" rule is OFF, or it blocks the answer crawlers before robots.txt.)

# ─────────────────────────────────────────────────────────────────────────────
# AI crawler policy — AC 2026-08-17
#
# Everything above blocked every AI agent, which was correct while there was
# nothing here to find. Now that the content is off Shopify and consolidating,
# the stance follows the Content-Signal already declared at the top of this
# file: search=yes, ai-train=no, use=reference.
#
# The split that matters is RETRIEVAL vs TRAINING:
#
#   RETRIEVAL agents fetch a page to answer someone's question and cite the
#   source. Blocking them is what makes a brand invisible in AI answers. These
#   are ALLOWED — deliberately by omission, see the note below.
#
#   TRAINING crawlers ingest the corpus into model weights. Nothing comes back,
#   and the content cannot be withdrawn later. These stay blocked.
#
# ⚠️ WHY THE ALLOWED AGENTS HAVE NO ENTRY HERE. robots.txt groups do NOT
# inherit. Giving OAI-SearchBot its own "Allow: /" group would detach it from
# the User-agent:* group above and hand it /admin, /finance, /deck and every
# other private surface. Allowed agents must fall through to * so the internal
# Disallow list still applies to them. Do not "helpfully" add groups for them.
#
# Allowed via *: OAI-SearchBot · ChatGPT-User · Claude-User · Claude-SearchBot
#                PerplexityBot · Googlebot · Bingbot · DuckDuckBot
#
# Blocked below: pure training crawlers.

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: anthropic-ai
Disallow: /

User-agent: CCBot
Disallow: /

User-agent: Bytespider
Disallow: /

User-agent: meta-externalagent
Disallow: /

User-agent: Amazonbot
Disallow: /

# Two judgment calls, each one line to flip:
#
# Google-Extended gates Gemini grounding AND Gemini training — Google does not
# separate them. Blocking it keeps the corpus out of training but also out of
# Gemini's answers. It is currently ALLOWED (no entry), because being cited is
# the point of this exercise. Add a Disallow group to reverse.
#
# Applebot-Extended is Apple Intelligence training. Currently ALLOWED; Apple is
# also the platform the app ships on. Add a Disallow group to reverse.

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