Glossary AI

MCP (Model Context Protocol)

Also known as: Model Context Protocol

An open protocol created by Anthropic that lets AI models connect to external tools and data sources. The standard for the tool layer of the AI coordination stack. MCP servers wrap a tool (a database, an API, a filesystem) and expose it to any MCP-compatible model.

Related terms

A2A (Agent-to-Agent Protocol)

A protocol that lets AI agents talk directly to each other. If one agent needs help from another, A2A is the language they use ...

MCP Server

A program that wraps an external tool or data source and makes it accessible through MCP. OTP ships an MCP server (otp-mcp-serv...

Organization Transport Protocol (OTP)

The protocol and platform for publishing, comparing, and learning from organizational coordination intelligence. Operates above...

The Three-Layer AI Coordination Stack

AI coordination operates at three layers: Tool (MCP — how agents call tools), Agent (A2A — how agents talk to each other), and ...

Build with this on OTP

If this term shows up in your playbook, it belongs in your OOS.

OTP encodes coordination intelligence so teams of humans and AI agents can run on it.

Get started Browse practices All terms →

Found an issue with this definition? Tell us and we'll fix it.