# Crawl us. Discovery is the point - but crawl the pages, not the paywall. # # Measured 2026-08-07 over 24h: Amazonbot spent 42% of its fetches on /v1/ # URLs that answer 402 and can never be indexed; Meta 34%, GoogleOther 34%. # Googlebot figured this out on its own (1%). The disallow below points # everyone else's budget at the HTML pages, which is where the content is. # Everything under /v1/ is documented in the machine entry points below - # an agent that wants the API does not need to crawl it blind. # The exceptions are not API: /v1/answers/* and /v1/vat/determine serve # HTML documentation pages and they are meant to be indexed. Longest # match wins for Google and Bing, so these Allow rules beat Disallow. # /provenance/{cc}/{series} dumps are source fingerprints for machines, # not search results. Disallow them so crawl budget goes to /{cc}/{series}. User-agent: * Disallow: /v1/ Disallow: /provenance/ Allow: /v1/changes Allow: /v1/calendar Allow: /v1/business-days/ Allow: /v1/answers/ Allow: /v1/vat/determine # Google and Bing get the same treatment as everyone else. They used to # have a blanket Allow because both already self-manage their /v1/ budget # sensibly - Googlebot spends about 1% there. Closed anyway: Search Console # reports 130 pages 'Discovered - currently not indexed', which is Google # declining to spend crawls, and every crawl of a 402 is one not spent on a # page that can actually be indexed. User-agent: Googlebot Disallow: /v1/ Disallow: /provenance/ Allow: /v1/changes Allow: /v1/calendar Allow: /v1/business-days/ Allow: /v1/answers/ Allow: /v1/vat/determine User-agent: Bingbot Disallow: /v1/ Disallow: /provenance/ Allow: /v1/changes Allow: /v1/calendar Allow: /v1/business-days/ Allow: /v1/answers/ Allow: /v1/vat/determine # Paid endpoints answer 402 (x402) to an unpaid request rather than 403 — # nothing is hidden; the disallow is crawl-budget hygiene, not secrecy. Sitemap: https://usaref.dev/sitemap.xml # Machine-readable entry points: # https://usaref.dev/llms.txt endpoints, prices, example bodies # https://usaref.dev/catalog every country and series held # https://usaref.dev/openapi.json OpenAPI description # https://usaref.dev/.well-known/x402.json x402 resource manifest # https://usaref.dev/v1/changes free change + freshness feed # https://usaref.dev/v1/calendar free re-verification calendar (not a gazette schedule) # https://usaref.dev/v1/business-days/ free business-day tools: is-business-day, add, between # Human-readable pages, one per country and per series: # https://usaref.dev/countries every country we hold # https://usaref.dev/{country}/{series} e.g. /za/vat — value, source, history