# SceneSat — https://scenesat.com/
#
# The archive is the point: shows, guest stations, artists, tracks and video recordings are all
# meant to be indexed and linkable. Only the surfaces with no standing value to a search result
# are excluded — endpoints that are per-visitor, machine-only, or a redirect hop.

User-agent: *

# JSON APIs: machine contracts for the site's own JS and the mobile apps, not pages.
Disallow: /api/

# Authentication round-trips. Never useful in an index, and crawling them creates sessions.
Disallow: /auth/

# Per-visitor surfaces: browser-local preferences and the legacy-account linking flow.
Disallow: /settings
Disallow: /account/

# Byte delivery rather than documents — the show/track PAGES that link these stay indexable.
#
# These MUST be written as the paths this site actually serves. `Disallow:` is a prefix match, so
# a rule only excludes what literally starts with it. An earlier version of this file listed
# /download/, /listen/, /peaks/ and /waveform/ — Korolev's route table, not this site's — and
# excluded nothing at all here, because the audio lives at /show/{id}/download. The result,
# measured over the access log on 2026-08-31: 3,722 archive downloads by GPTBot (1,736),
# DataForSeoBot (995) and ClaudeBot (991), every one of which honours this file and would have
# stayed away had it been asked. Verify a new rule against `$router->get(...)` in public/index.php
# before trusting it.
Disallow: /media/
Disallow: /show/*/download
Disallow: /track/*/download

# The guest-station relay. A crawler here is worse than a wasted request: the connection is a live
# audio stream held open for as long as the "listener" stays, and it occupies a php-fpm worker for
# the whole of it. Nothing indexable is behind it.
Disallow: /guest-stations/*/stream

# Legacy short-link that only 302s to the real watch page; the target is indexable.
Disallow: /videoarchive/

# Everything else is fair game, explicitly — this file must never carry a blanket noindex.
Allow: /

# What exists, rather than only what is reachable by following links. This matters more here
# than usual: the apex served a 404 for three weeks before the 2026-08-30 cutover, so a crawler's
# picture of this site is stale. The index points at per-section children under /sitemap/.
Sitemap: https://scenesat.com/sitemap.xml