# Quotle.info — verified quote provenance. Crawlers and AI agents welcome. # For agents: llms.txt maps the site; the quote-verification API is described at # /openapi.json and /.well-known/ai-plugin.json (verify a quote or a whole deck). User-agent: * Allow: / # Editorial working files. These are build INPUTS, not site content — nothing on the site links to # them and nothing reads them at runtime. harvest-queue.json in particular carries the unpublished # candidate backlog and, necessarily, the full text of candidates we DECLINED to publish under the # hate/harm skip bar (the dedup that stops them being re-harvested matches on that text, so it # cannot simply be removed). Keeping them out of search indexes is the point here. # NOTE: this stops crawling, not direct access — the file is still served, because GitHub Pages # with .nojekyll serves everything committed. The structural fix is to publish Pages from a # subdirectory so working files are not in the web root at all. Disallow: /data/harvest-queue.json Disallow: /data/harvest-queue.md Disallow: /data/harvest-config.json Disallow: /data/harvest-config.example.json Disallow: /data/song-queue.json Disallow: /data/demand-cache.json # TWO SITEMAP URLS, IDENTICAL BYTES — this is deliberate, do not "tidy" it to one. # # Google's Search Console entry for /sitemap.xml entered a stuck "Couldn't fetch" state on # 2026-07-18 (Type "Unknown", Last read empty, 0 discovered) and NEVER recovered. It was not the # file: on 2026-07-25 the identical bytes at /sitemap-full.xml were read on the first attempt — # Status Success, 2,048 pages discovered — as was a 258-byte 3-URL control at /sitemap-test.xml. # Bing read /sitemap.xml successfully throughout. The failure was per-ENTRY state on that one URL, # and resubmitting the same URL never clears it: the row's menu offers no "remove" (there is # nothing successfully fetched to remove), so the dead entry is permanent and un-deletable. # # /sitemap-full.xml is therefore the URL Google actually reads. /sitemap.xml stays listed because # Bing and other crawlers read it fine and it is the conventional location. Sitemap: https://quotle.info/sitemap-full.xml Sitemap: https://quotle.info/sitemap.xml