AI Resources

Machine-readable / AI-friendly resources for discovery and indexing. Public read-only unless otherwise stated.

Manifest: /ai/manifest.json | /ai/manifest.toon | /ai/openapi.json

LLM guide: /llms.txt | /llms-full.txt

Recommended for AI agents

Start: /ai/manifest.json
Incremental sync: /ai/feeds/updates.jsonl?since=...
Search/detail: /ai/search.json?q=..., /ai/products/{sku}.json
Compact LLM mirrors: /ai/toon and /ai/toon/meta.toon

JSON / CSV Feeds

Updates feed supports optional ?since=UNIX_TIMESTAMP. List feeds may support ?page=, ?limit=, and ?since=. Use /ai/feeds/meta.json for source, freshness, pagination, and mirror metadata.

{"id":"ABC123","sku":"ABC123","title":"Example Product","summary":"Short description","url":"https://seppleaf.com/products/abc123"}

Details + Search

/ai/search.json?q=gold
/ai/search.toon?q=gold
/ai/products/{sku}.json
/ai/products/{sku}.toon
/ai/guides/{slug}.json
/ai/guides/{slug}.toon
/ai/glossary/{slug}.json
/ai/glossary/{slug}.toon
/ai/media/{id}.json
/ai/media/{id}.toon

{"query":"gold","limit":20,"items":[{"id":"product:ABC123","type":"product","title":"Example Product"}]}

Schemas

Data Dictionary

id — Internal content or product identifier.
sku — SKU / product code when applicable.
type — Content type, endpoint type, or search result type.
title/name — Human-readable title or name.
summary/description — Short or long description.
url — Canonical public URL.
uri — Stable route slug/URI.
brand/category/tags — Catalog navigation and grouping fields.
last_updated/uploaded_at — UNIX timestamp when present.
mime_type/pdf_url — Media/document metadata when present.
image_url/price/availability/product_type/material/finish/size/related_products/documents/charts/breadcrumbs — Best-effort product enrichment fields; chart/SDS/TDS/PDF data is MVP until the overhaul is complete.

MCP

MCP endpoints support page, limit, and since query params where applicable. Blog MCP routes may require API-key access depending on endpoint/config; open feed mirrors are listed above.

MCP is a public, read-only content gateway that exposes curated site data (products, guides, glossary, media) via MCP JSON-RPC tools and simple REST lists. It provides AI-friendly, scoped access to specific content domains with stable schemas because AI systems need structured, selective access: Precise, predictable fields (IDs, titles, summaries, URLs, tags) made easy for tools to search and fetch without scraping HTML.

{"page":1,"limit":50,"count":1,"items":[{"id":"ABC123","title":"Example Product"}]}

Product detail pages include structured data for key product fields.

Sitemaps

TOON

feed: products\nitems[1]{id,sku,title,url}:\n  ABC123,ABC123,"Example Product",https://seppleaf.com/products/abc123

TOON (Token-Oriented Object Notation) is a compact, human-readable encoding of JSON-style data for AI systems. These TOON endpoints are LLM-oriented mirrors of the JSON/feed data, with lower token overhead and explicit structure. Use JSON/JSONL/CSV/MCP as canonical machine sources, and TOON when prompt readability/token budget matters. Product, update, guide, glossary, media, document, chart, and blog mirrors are available; document/chart/SDS/TDS/PDF fields are MVP until the data overhaul is complete. See also https://axi.md

AI Resources

Contact Us

General

Social