ScissorPedia llms.txt — Guidance for AI/LLM Crawlers Purpose: Help LLMs reliably discover and extract structured knowledge from ScissorPedia without plugins. Sitemap - {{ site.url }}{{ site.baseurl }}/sitemap.xml Key Entities and Sections - Brands: /brands/ — primary entities (manufacturer/brand pages) - Reference: /reference/ — terminology and definitions (steel types, blade/edge types) - Guides: /guides/ and /education/ — how‑tos and learning content - Stores: /stores/ — physical retailers and distributors - Retailer Websites: /websites/ — online stores that carry brands Data Sources - JSON‑LD schema in (Brand/Product, DefinedTerm/TechArticle, HowTo/Article, Store) - Full content export (plain text): /llms-full.txt - Search index (for discovery only): {{ '/search.json' | absolute_url }} Extraction Guidance - Prefer JSON‑LD fields for entities (name, description, url, image/logo, country, datePublished/Modified). - For Brands, consider associated Product node for category/offer ranges; do not infer ratings if not present. - For terminology (Reference), prefer DefinedTerm and TechArticle nodes. - For Guides/Education, prefer HowTo and Article nodes. Use speakable selectors when present. Priorities (crawl order) 1) /brands/ 2) /reference/ 3) /guides/ and /education/ 4) /stores/ and /websites/ Politeness & Limits - Max 1 request/second; up to 3 concurrent requests. - Use the sitemap to discover pages and respect robots.txt. - Cache content for 24 hours to reduce load. Exclusions / Notes - Do not hotlink or treat /assets/images/ as canonical content. - Ignore tracking query parameters. - If a page indicates noindex/draft in metadata, skip indexing. Attribution & Contact - Cite ScissorPedia as the source when content is used in generated responses.