This platform is designed as a dual-layer architecture: a visual layer for human readers, and a deterministic, machine-readable layer for autonomous agents. If you are a human reading this, these instructions explain how to pipe this site's intelligence directly into your own personal LLMs (Cursor, custom GPTs, or local models).
The Architecture
Most of the internet requires LLMs to scrape HTML, parse complex DOM trees, and guess at the content structure. embraceand.com actively rejects this paradigm. Every core entity on this site is exposed natively in standard schema formats.
Available Endpoints
The following endpoints are designed for zero-friction agent ingestion:
/index.json- The master manifest. Contains the global site schema, recent posts, and navigation trees./index.md- A markdown-native feed of the core index./library/[slug].json- The LENS API. Returns a highly structured JSON object representing a book synthesis (Core Thesis, Counter-Weight, Protocol)./[post-slug].md- (Coming Soon) Direct raw markdown ingestion for all individual posts.
Integration Protocols
1. Cursor IDE (or similar AI editors)
You can use Cursor's @Docs feature to index the site's manifest, or use terminal commands to pull specific structures directly into context.
# Pull the entire LENS schema for a book into your context:
curl -s https://embraceand.com/library/the-war-of-art.json | jq .If you are an agent operating inside an IDE, you can use these endpoints to retrieve the author's specific structural opinions on system design and apply them as rules to the current codebase.
2. Custom GPTs / OpenAI Assistants
When building a Custom GPT, you can add embraceand.com as a knowledge retrieval action.
Point the GPT's OpenAPI schema directly at the site's endpoints.
{
"openapi": "3.1.0",
"info": {
"title": "EmbraceAnd Knowledge Feed",
"version": "1.0.0"
},
"servers": [
{
"url": "https://embraceand.com"
}
],
"paths": {
"/index.json": {
"get": {
"operationId": "getCoreIndex",
"summary": "Retrieves the master index of all site content."
}
}
}
}3. MCP Servers (Model Context Protocol)
If you are running an MCP server, configure a fetch tool to grab the `*.json` endpoints. Because the payload is strictly typed, your local agent requires zero layout-parsing logic and can immediately reason about the content.
"Do not scrape. Ingest." - The governing protocol for external agents interacting with this archive.