PositioningIndexBot -- crawler disclosure
============================================================

User-Agent
  PositioningIndexBot/1.0 (+https://github.com/rushogen/positioning-index#crawling-policy; marketing-page positioning research; one request per page per day; honours robots.txt)

What it does
  Fetches the publicly accessible homepage and pricing page of the B2B SaaS
  companies listed at https://github.com/rushogen/positioning-index#crawling-policy, and records a small number of short
  factual strings from them: the hero headline and subhead, the category noun,
  the meta title and description, published pricing tier names and prices,
  customer logo names, and quantified marketing claims.

  The purpose is to observe how software companies change their own positioning
  over time. Nothing is republished at length; the index stores short excerpts
  and links back to the source page.

Rate
  At most one request per page per day. Requests are strictly serial, and two
  requests to the same host are separated by at least 60 seconds, or by the
  host's Crawl-delay where that is longer. robots.txt is fetched at most once
  per host per run.

  The crawler is not always-on. It runs when a person or a manually triggered
  workflow asks it to, and then the process exits.

What it honours
  robots.txt per RFC 9309, including per-agent groups, longest-match rule
  precedence, and Crawl-delay.
  Content-Signal declarations (content-signals.org). This crawler indexes; it
  does not train models and does not supply a generative system.
  HTTP 429 and Retry-After.
  If a robots.txt cannot be fetched, or returns 5xx, we do not crawl.

How to stop it
  Add this to your robots.txt:

      User-agent: PositioningIndexBot
      Disallow: /

  It takes effect within 24 hours, which is the robots.txt cache lifetime.
  Or open an issue at https://github.com/rushogen/positioning-index#crawling-policy and the domain will be removed from the
  seed list entirely.

Not collected
  No personal data. No authenticated pages. No form submissions. No pages
  outside the two URLs published per company in seed/companies.json.