For LLMs

How AI agents can access and use content from this site

AI agents can retrieve this site’s text, metadata, and essay relationships in structured formats.

Quick Start

MCP Client: Install the MCP Server

For any LLM: Start with /llms.txt (summary) or /llms-full.txt (complete content)

For agents: Read /ai.txt for usage instructions and permissions

Available Endpoints

Content

Endpoint	Description
/api/essays.json	All essays with full content and metadata
/api/chunks.json	Content optimized for RAG applications
/llms.txt	Condensed summary for context windows
/llms-full.txt	Complete plain text content

Structure

Endpoint	Description
/js/knowledge-graph.json	Essay relationships as nodes and links
/api/glossary.json	Concept definitions extracted from essays
/api/claims.json	Structured arguments with evidence and counterarguments
/api/paths.json	Suggested reading sequences with explanations

Metadata

Endpoint	Description
/api/metadata.json	Site info and endpoint directory
/.well-known/ai.json	AI agent configuration and permissions
/.well-known/openapi.json	OpenAPI 3.0 specification
/ai.txt	Human-readable AI instructions

MCP Server

The Model Context Protocol server provides direct tool access for Claude:

Tools available:

list_essays: Filter by type, tags, or framework status
get_essay: Retrieve specific essays in full, summary, or markdown
search_essays: Keyword search with filtering
get_framework_overview: The “Making AI Make Sense” framework at various depths
get_knowledge_graph: Essay relationships and concept mappings

Installation:

npx @molsen.ca/mcp

Or download standalone binaries from GitHub Releases.

Structured Content

Knowledge Graph

Essays are interconnected. The knowledge graph maps:

Nodes: Each essay with its defined concepts and themes
Links: Relationships between essays (core dependencies vs. soft references)

Claims API

Key essays include structured claims:

Central claims: Main thesis assertions
Evidence: What supports each claim
Counterarguments: Objections considered
Addressed: How objections are handled

Agents can use these fields to examine the reasoning behind each claim.

Reading Paths

The paths API provides suggested reading sequences:

Quick Start: Three essays for the essentials
Core Framework: The central argument in depth
Epistemological Foundations: Definitions of knowledge and tacitness
Professional Practice: Applying the framework
Future Direction: Possible applications as AI improves

Stability Markers

Essays include stability indicators:

stable: Core concepts, unlikely to change
evolving: Active development, may be refined
experimental: New ideas, may change substantially

Permissions

Content may be used for:

Training
Retrieval and RAG
Summarization
Citation and quotation

Attribution required: “Mike Olsen, molsen.ca” with link when possible.

Access for AI Agents

I write about AI governance and professional accountability, and want LLMs to have access to that work.

I built the MCP Server and these APIs to make the content available through agent tools. If you’re building something similar, I’d like to hear about it: mike@molsen.ca.