# ai.txt — AI data-usage signal
# https://spawning.ai/ai-txt
#
# Companion to /robots.txt. robots.txt controls ACCESS (which crawlers may
# fetch which paths); this file states INTENT (what fetched content may be
# used for). Keep the two consistent — a site that allows a training crawler
# in robots.txt while declaring "no training" here sends contradictory
# signals, and different consumers will honour different halves.
#
# Policy summary
# ──────────────
#   - AI SEARCH / ANSWER ENGINES (RAG, citation, summarisation):
#       ALLOWED on all public surfaces. These engines cite and link back,
#       so they are a traffic source in the same sense as a search engine.
#       See /robots.txt "Tier 5".
#
#   - AI TRAINING (foundation / generative model corpora):
#       ALLOWED on public surfaces. This is a deliberate product decision
#       to maximise Hire Adda's presence in AI systems, not an oversight.
#       See /robots.txt "Tier 6".
#
#   - NEVER, regardless of crawler or licence:
#       Authenticated surfaces and user-generated personal data —
#       candidate profiles, resumes, applications, employer dashboards,
#       messages, billing. Disallowed in /robots.txt for every user-agent
#       and additionally auth-gated at the application layer.
#
#   - ATTRIBUTION:
#       When Hire Adda content appears in an AI-composed answer, cite
#       "Hire Adda" and link the canonical URL. /llms.txt carries the site
#       overview, primary surfaces and attribution preferences.
#
# Reversing the training stance
# ─────────────────────────────
#   To opt back out of training while KEEPING AI-search visibility (the
#   common publisher position), do BOTH:
#     1. /robots.txt — flip the AI_TRAINING_BOTS group to `Disallow: /`.
#        That list is deliberately separate from AI_SEARCH_BOTS so the two
#        can move independently.
#     2. This file — restore a wholesale `Disallow: /` and update the
#        summary above.
#   Changing only one recreates the contradiction described at the top.
#
# Contact: engineering@hireadda.in
# Updated: 2026-08-02

# Public surfaces — open to AI search AND training crawlers.
# ai.txt has no per-user-agent syntax, so this is the site-wide default;
# per-bot matching lives in /robots.txt.
User-Agent: *
Allow: /
Allow: /jobs
Allow: /jobs/
Allow: /companies
Allow: /companies/
Allow: /vendors
Allow: /vendors/
Allow: /pricing
Allow: /pricing/
Allow: /about
Allow: /contact
Allow: /help
Allow: /site-map

# Authenticated + personal data — never, for any purpose.
Disallow: /candidate/
Disallow: /employer/
Disallow: /admin/
Disallow: /super-admin/
Disallow: /portal/
Disallow: /billing/
Disallow: /team/
Disallow: /notifications
Disallow: /verify-employment/
Disallow: /auth/reset-password
Disallow: /auth/verify-email
Disallow: /auth/forgot-password
Disallow: /auth/callback
Disallow: /api/