Home
Back to Glossary

MCP

Pages
7
References
6
Related Terms
8

Definition

MCP is an open protocol for connecting LLM applications to external tools, data sources, prompts, and contextual resources.

Background

Introduced by Anthropic in 2024, Model Context Protocol reframes integrations around a common client-server interface instead of bespoke tool connectors for each AI application.

Position

It sits between AI agents, tool access, authentication, and memory systems. Knowledge Graph and Graphiti are adjacent knowledge backends; OAuth, OIDC, and PKCE are adjacent authorization concepts.

Distinctions

  • It is distinct from other MCP expansions such as Microsoft Certified Professional.
  • MCP is not a model; it is a connection protocol used by model-based applications.
External Reference Model Context Protocol specification Official

Primary source-backed reference selected for this concept.

A concept map for MCP.

Page Context

  • AI/LLM/Ontology/Organizational Memory
    1. Executive Summary Rather than being a "knowledgeable entity," modern LLMs are probabilistic pattern generators that learn from large volumes of language, code, images, and be...
    Quote: AI/LLM/Ontology/Organizational Memory

    ai-systems

  • MCP and Agent Skills for Browser E2E Testing
    MCP and Agent Skills for Browser E2E Testing 1. Executive Summary As of May 2026, the right way to let an AI agent run browser E2E checks is not simply "add a browser MCP server...
    Quote: MCP and Agent Skills for Browser E2E Testing

    developer-tools

  • AI agent memory platform created with Graphiti and MCP
    1. Executive Summary By combining Graphiti and MCP, AI agents can be provided with "memory that persists across conversation sessions" and "a standard point of contact from exte...
    Quote: AI agent memory platform created with Graphiti and MCP

    ai-systems

  • How LLM Training, Fine-Tuning, RAG, and Agents Differ
    How LLM Training, Fine-Tuning, RAG, and Agents Differ 1. Executive Summary The most common mistake in LLM discussions is to treat pretraining, fine-tuning, prompting, RAG, tool ...
    Quote: How LLM Training, Fine-Tuning, RAG, and Agents Differ

    ai-systems

  • OAuth 2.1 PKCE Flow and MCP Authentication Authorization Practical Guide
    OAuth 2.1 PKCE flow and MCP authentication authorization practical guide 1. Executive Summary In practice, MCP server authentication and authorization can be easily understood a...
    Quote: OAuth 2.1 PKCE Flow and MCP Authentication Authorization Practical Guide

    developer-tools

  • Research Log: LLM Training, Fine-Tuning, RAG, and Agents
    Research Log: LLM Training, Fine-Tuning, RAG, and Agents Environment - model: gpt-5.4-mini - skill: research-report - prompt source: ops/codex/prompts/daily-issue-research.md Re...
    Quote: Research Log: LLM Training, Fine-Tuning, RAG, and Agents

    ai-systems

Pages