# llms.txt — Guidance for Large Language Model (LLM) crawlers
# Place at: https://education-matters-6de5fc.webflow.io/llms.txt
# Purpose: Declare how AI/LLM services may use site content.
# This is a human-readable, machine-parseable manifest (YAML-like).

version: 1.0
site: https://education-matters-6de5fc.webflow.io
owner: Education Matters Africa
contact: /contact
sitemap: /sitemap.xml

permissions:
  indexing: allow                 # OK to index pages for search / answers
  training: disallow             # Do NOT use content for model training
  summarization: allow           # OK to summarize pages on demand
  qa: allow                      # OK to answer questions with brief excerpts
  embedding: allow               # OK to create vector embeddings for retrieval
  derivative_works: require-permission
  commercial_use: require-permission

rate_limits:
  crawl_delay_seconds: 10        # Be polite
  cache_ttl_seconds: 604800      # 7 days

content_rules:
  images: allow-nontraining      # May display/thumbnail, but not train on them
  documents: allow               # PDFs and docs may be read for on-demand use
  datasets: disallow             # No bulk copying
  personal_data: disallow        # Never collect personally identifiable info
  login_pages: disallow

scopes:
  # If you have different subpaths on production, update these.
  allow:
    - /                          # Home
    - /about-us                  # About (update if slug differs)
    - /usap-community-school     # USAP overview
    - /programs                  # Programs hub
    - /programs/capstone-fellowship
    - /programs/zimbabwe-career-connect
    - /programs/zimpact
    - /partners                  # Program partners (adjust if needed)
    - /news                      # News/Blog index
    - /usap-25                   # USAP @ 25
    - /donate                    # Donate
    - /contact                   # Contact
  disallow:
    - /admin
    - /search
    - /login
    - /?*                        # Query-only pages

attribution:
  required: true
  style: link-to-canonical       # Link back to the original page when quoting
  quote_limit_words: 75

legal:
  license: "All rights reserved unless otherwise noted."
  notice: "Use of site content is subject to these terms; contact us for broader rights."

notes:
  - "This file complements robots.txt LLM user-agent rules (e.g., GPTBot, Google-Extended, ClaudeBot, PerplexityBot)."
  - "If production domain differs, duplicate this file there and update 'site:'."