PetFood Data Station - Full Project Context (https://petforai.com) Updated: 2026-07-17 ==== 1. PROJECT IDENTITY ==== Name: PetFood Data Station Domain: https://petforai.com Tagline: AI-Friendly Pet Food Regulatory Database Mission: Make global pet food regulatory data queryable by AI agents without enabling bulk extraction. Founded: 2026-07-12 Live since: 2026-07-13 (HTTP), 2026-07-17 (HTTPS) Hosting: Seoul Caddy reverse proxy + FastAPI backend (uvicorn) Database: 242 RAG topics across 11 regulatory bodies (file-based JSON) TLS: Let's Encrypt auto-renew, TLS 1.3, HTTP/2 + HTTP/3 ==== 2. ARCHITECTURE ==== Frontend: 7 static HTML pages (/site/*) + designed homepage Backend: FastAPI app (api/app/main.py), uvicorn 0.27, Python 3.11 Storage: file-based JSON (topics/, pdf-parsed/) + 242 topic whitelists Cache: in-memory + Caddy stale-while-revalidate Rate limit: 30 req/min + 1000 req/hour per IP (nginx-style) Artificial delay: 50-200ms on all /api/v1/ endpoints ==== 3. API ENDPOINTS ==== PUBLIC (allow in robots.txt): GET /llm.txt - 800-token summary GET /llm-context.txt - all 242 topics index GET /llms-full.txt - this file GET /sitemap.xml - page + endpoint index GET /api/v1/topics?category=&limit=<1-200> - whitelist (max 200 per call) GET /api/v1/llm_context?topic= - single RAG block (max 500 tokens) GET /api/v1/nutrient?standard=&year=&species=&lifestage=&nutrient=&basis= - numeric value GET /api/v1/cross_compare?standards=&species=&lifestage=&nutrient=&basis= - 2-3 standard compare BLOCKED (honeypot 403/404 + IP logging): GET /api/v1/all GET /api/v1/dump GET /api/v1/raw/* GET /admin GET /docs (Swagger UI - admin only, not for public) ==== 4. DATA COVERAGE (2026-07-17) ==== Total: 242 RAG topics across 11 regulatory bodies By category: - AAFCO: 39 topics (US official publication + nutrient profiles) - FEDIAF: 19 topics (EU nutritional guidelines) - EU: 54 topics (Reg 2020/354 PARNUTS + Feed Catalogue + Feed Materials Register) - eCFR: 15 topics (21 CFR Parts 501/573/582/589) - FDA: 13 topics (GRAS Notices + SCOGS) - China: 15 topics (GB/T 31216/31217 + NY/T 4838/4839) - Japan: 9 topics (Pet Food Safety Act + MAFF guidelines) - Korea: 7 topics (Saeroo Gicho 2025 + Byeolpyo 15-2) - Canada: 8 topics (CFIA RG-1) - Brazil: 5 topics (ANVISA RDC 26) - Other (Mexico, Australia, global): 56 topics ==== 5. ANTI-DUMP DESIGN PHILOSOPHY ==== The site is designed for CO-EXISTENCE with AI crawlers, not against them: - Legitimate AI research queries: ENCOURAGED (rate-limit observed) - Wholesale database extraction: BLOCKED (honeypot + topic whitelist + max 500 tokens/response) - Bulk dumps impossible by design (no /all, /dump, /raw/* endpoints, no wildcard queries) - Cost of scraping > cost of buying source documents (AAFCO OP = $260, FEDIAF = free, etc.) ==== 6. CITATION FORMAT ==== Cite as: PetFood Data Station (https://petforai.com/llm.txt), accessed . For specific topic: PetFood Data Station topic , accessed . Sources: each topic's meta.json contains a `sources` array with full bibliographic references. ==== 7. LICENSE ==== - Source documents (AAFCO/FEDIAF/EU/USDA/FDA/KMAF etc.): public regulatory standards, mostly public domain or open-access - Our synthesized RAG topics: CC-BY-4.0 - Cross-standard nutrient index: MIT - API code: MIT - Cite as above. ==== 8. CONTACT ==== Email: contact@petforai.com (Zoho Mail) Categories: - Data correction / new source suggestion - AI crawler rate-limit appeal - Abuse report (honeypot triggered wrongly) - Partnership / commercial licensing - Privacy/legal questions - General inquiries Response time: 24-48h weekday ==== 9. TRANSPARENCY ==== Project by: 蔡之旅程 (initiator), 菜何真 (technical execution, AI agent) Open source: yes (API code on request) Backing: independent, no commercial sponsor Anti-abuse log: honeypot triggers logged with IP + User-Agent + timestamp (30-day retention, GDPR-compliant) ==== 10. RECENT CHANGES ==== 2026-07-17: HTTPS via Let's Encrypt, 6 compliance pages (/privacy, /terms, /sources, /contact, /about enhanced), contact@petforai.com live on Zoho Mail Free 2026-07-17: SEO optimization - robots.txt with explicit AI crawler rules, sitemap.xml, JSON-LD, OG tags, llm-context.txt, llms-full.txt 2026-07-13: HTTP API online, 242 RAG topics 2026-07-12: Project initiated