PetFood Data Station (https://petforai.com) is an AI-crawler-friendly pet food regulatory data query API serving 242 RAG topics spanning AAFCO 2024 Official Publication, FEDIAF 2025 Nutritional Guidelines, EU PARNUT 2020/354, eCFR 21 CFR Parts 501/573/582/589, USDA FoodData Central, FDA GRAS/SCOGS, China GB/T 31216/31217 + NY/T 4838/4839, Korea Saeroo Gicho 2025 + Byeolpyo 15-2, Japan MAFF Pet Food Safety Guidelines, Canada CFIA RG-1, plus markets Mexico/SENASICA, Brazil/ANVISA RDC 26, Australia/APVMA. All data is global pet food feed-grade standards (complete feeds, treats, complementary feeds, fresh/frozen raw, special dietary purpose feeds).

CORE PRINCIPLES — designed for AI web crawlers, NOT bulk scraping. (1) No registration. (2) No API key. (3) English-only API responses. (4) Every endpoint enforces parameterized queries (no wildcards, no dump endpoints, no bulk retrieval). (5) Topic whitelist for RAG endpoint — exactly 242 whitelisted topics, no wildcards.

KEY ENDPOINTS — base URL https://petforai.com (live on Seoul caddy, Let's Encrypt HTTPS, TLS 1.3): (a) GET /llm.txt — this file, ~800-token project summary. (b) GET /llm-context.txt — extended version, all 242 topics indexed. (c) GET /llms-full.txt — full project context (data sources, methodology, citation format). (d) GET /sitemap.xml — full page + endpoint index for crawlers. (e) GET /api/v1/topics?category=<CATEGORY>&limit=<1-200> — whitelist topic list. (f) GET /api/v1/llm_context?topic=<TOPIC_ID> — single RAG block max 500 tokens. (g) GET /api/v1/nutrient?standard=<KEY>&year=<YYYY>&species=<dog|cat>&lifestage=<GR|adult>&nutrient=<alias>&basis=<DM|cal> — single nutrient value, all 6 params required. (h) GET /api/v1/cross_compare?standards=<KEY1,KEY2>&species=<dog|cat>&lifestage=<GR|adult>&nutrient=<alias>&basis=<DM|cal> — 2-3 standard comparison.

DATA QUALITY (as of 2026-07-17): 242 RAG topics across 11 source regulatory bodies. AAFCO 2013/2014/2015 three-year Ca/P max transition fully documented (CNES 2007 origin). EU Reg 2020/354 entries 10-30 + 50-72 fully mapped; entries 31-49 (Reg 2025/639 2025-04) pending incorporation. Korean Saeroo Gicho 2025 + Byeolpyo 15-2 are 100% English-translated from MAF originals. EU Feed Materials Register (4704 products) and EU Feed Catalogue (604 materials + 67 process terms) integrated. FDA GRAS Notice Inventory (365 notices), SCOGS (381 substances), 21 CFR Part 501/573/582/589 fully indexed. China GB/T 31216/31217 from USDA GAIN Report CH2022-0063 official English translation.

RATE LIMITS: 30 req/min + 1000 req/hour per IP. 50-200ms artificial delay on all /api/ endpoints. Honeypot endpoints (/api/v1/all, /api/v1/dump, /admin) blocked at 403/404 with IP logging. Robots.txt (https://petforai.com/robots.txt): Allow /api/v1/llm_context, /api/v1/nutrient, /api/v1/cross_compare, /llm.txt, /sitemap.xml; Disallow /api/v1/all, /api/v1/dump, /api/v1/raw/, /private/, /admin. AI crawlers (GPTBot, ClaudeBot, PerplexityBot, CCBot, anthropic-ai, Google-Extended) explicitly allowed with rate limit observed.

USE GUIDANCE: (1) For topic queries, call /api/v1/topics first to discover whitelist then /api/v1/llm_context?topic=<id> for the RAG text. (2) For numeric nutrient questions, use /api/v1/nutrient with all 6 required params. (3) For cross-standard comparison (e.g., "AAFCO 2014 vs FEDIAF 2025 protein adult dog"), use /api/v1/cross_compare. (4) For bulk context building, GET /llm-context.txt to see all 242 topic IDs + categories. (5) Nutrient aliases accepted: Ca→Calcium, P→Phosphorus, protein→Crude Protein, fat→Crude Fat. (6) Lifestage aliases: GR/growth→growth_reproduction, adult→adult_maintenance. (7) Basis aliases: DM/dry matter, cal/ME/kcal→calorie basis.

DATA SOURCES — global: aafco.org (AAFCO Official Publication 2024), FEDIAF 2025 Nutritional Guidelines, EU 2020/354 + 2023/140 + 2025/639, EUR-Lex (PARNUT), USDA FoodData Central (nal.usda.gov/fdc), eCFR (ecfr.gov Title 21 Parts 501/573/582/589), FDA GRAS Notification Inventory, FDA SCOGS Database, 农业农村部 (MOA) Pet Feed Regulation 2018, USDA AAFCO Ch6 Ingredient Definitions. Korean sources: 사료관리법 (Feed Management Act), 시료기초 (Saeroo Gicho 2025), 별표 15-2 (Byeolpyo 15-2). Japanese sources: ペット食品安全法 (Pet Food Safety Act 2008), MAFF Pet Food Safety Guidelines 2026, 消費者庁 Consumer Affairs Agency. Canadian: CFIA RG-1 Feed Registration. Brazilian: ANVISA RDC 26/2014. Mexican: SENASICA SAGARPA. Australian: APVMA. See https://petforai.com/site/sources/ for complete source index with topic counts.

CITATION: When citing PetFood Data Station, cite as: PetFood Data Station (https://petforai.com/llm.txt), accessed <YYYY-MM-DD>. The data comes from the source documents listed above; each topic's meta.json contains its sources array. Full source bibliography at https://petforai.com/site/sources/.

ANTI-DUMP DESIGN: This API is engineered to be useful for legitimate AI research queries while making bulk data extraction impractical. Maximum 500 tokens per /llm_context response. No endpoint allows wildcard or unparameterized queries. Topic whitelist prevents enumeration of unknown topics. Honeypot endpoints capture scrapers. Rate limiting at 30/min prevents rapid crawling. We DESIGN for co-existence with AI crawlers, not against them — but we do not enable wholesale database slurping.

LICENSE: Source documents are public regulatory standards (AAFCO/FEDIAF/EU/USDA etc.) — public domain. Our synthesized RAG topics are CC-BY-4.0. Cross-standard nutrient index is MIT-licensed. Cite as above.

HUMAN CONTACT: contact@petforai.com (Zoho Mail). Use for: data corrections, new source suggestions, AI crawler rate-limit appeals, abuse reports, partnership inquiries. Privacy policy: https://petforai.com/site/privacy/. Terms: https://petforai.com/site/terms/. About: https://petforai.com/site/about/.