# selda.ai — marketing site
#
# Kept in sync with what actually exists. A rule for a path that does not exist is
# noise that hides the rules that matter, so this file gets revisited whenever the
# page tree does. /compare came back on 2.9.2026 as six written pages, and
# /fi/oppaat and /fi/tapahtumat arrived with it; all three are crawlable and want
# to be, so they need no rule here, only the sitemap entry they have.
#
# The partnership proposals under /fi/kumppanuus are NOT listed as Disallow on
# purpose. They carry a noindex meta tag, and a crawler has to be able to fetch a
# page to read that tag: blocking them here would leave them eligible to appear in
# results as a bare URL, which is the opposite of the intent.
#
# ── ONE RULE OF THIS FILE THAT COST US SOMETHING ─────────────────────────────
#
# A named group REPLACES the `*` group for that crawler; the rules are not added
# together. Until 2.9.2026 the answer engines each had a group containing only
# `Allow: /`, which meant GPTBot, ClaudeBot and PerplexityBot were the only
# crawlers on the internet explicitly permitted to fetch `/for?…`, the per-lead
# preview addressed to one named company, and `/meet`. Every named group below
# now repeats the disallows. If a path is added to the `*` group, it has to be
# added to the others in the same edit.

User-agent: *
Allow: /
Disallow: /app/
Disallow: /api/

# The per-lead preview a prospect opens from a message. Each one is addressed to
# one company and has no business in an index.
Disallow: /for?
Disallow: /meet

# Selda cards. Each is a person's own public page, served through the /c/ rewrite.
Allow: /c/

# ── Social unfurlers ─────────────────────────────────────────────────────────
# These fetch one URL to render a preview card; they do not crawl. They are
# allowed everywhere on purpose, including /for? and /meet, because a link
# pasted into a chat should still show its card.

User-agent: Twitterbot
Allow: /

User-agent: facebookexternalhit
Allow: /

User-agent: LinkedInBot
Allow: /

# ── Answer engines ───────────────────────────────────────────────────────────
#
# Two different jobs, and both matter. Some of these build an index that a
# chatbot searches (OAI-SearchBot, PerplexityBot, Google-Extended); others fetch
# a page live because a person just asked a question that needs it (ChatGPT-User,
# Perplexity-User). Being absent from the second kind means never being cited in
# an answer, which is the modern version of not ranking.
#
# All of them read llms.txt as well, and it is kept current alongside this file.
# Every group repeats the disallows above, see the note at the top.

User-agent: GPTBot
User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: ClaudeBot
User-agent: Claude-Web
User-agent: anthropic-ai
User-agent: Claude-SearchBot
User-agent: PerplexityBot
User-agent: Perplexity-User
User-agent: Google-Extended
User-agent: Applebot-Extended
User-agent: meta-externalagent
User-agent: Amazonbot
User-agent: DuckAssistBot
User-agent: cohere-ai
User-agent: CCBot
User-agent: Bytespider
Allow: /
Allow: /c/
Disallow: /app/
Disallow: /api/
Disallow: /for?
Disallow: /meet

Sitemap: https://selda.ai/sitemap.xml
Sitemap: https://selda.ai/c/sitemap.xml