What is the Model Context Protocol (MCP)?

Home/Glossary/Model Context Protocol (MCP)

Architecture · Definition

Model Context Protocol (MCP)

The Model Context Protocol (MCP) is an open standard for connecting AI applications and agents to external tools, data sources, and systems. An MCP server exposes capabilities, such as searching a repository or querying a database, through a common interface, so any MCP-compatible agent can use them without a custom integration for each tool.

Also known as: MCP, MCP server

MCP and A2A

MCP connects an agent to tools and data; the Agent2Agent (A2A) protocol connects agents to each other. Enterprise multi-agent systems often use both: A2A for collaboration between agents, MCP for the systems each agent works with.

In the IBM stack

IBM Bob connects to enterprise tools through MCP, watsonx Orchestrate agents can use MCP-based tools, and AI Asset Discovery in watsonx.governance finds MCP servers so they can be brought under governance.

How Incede.ai approaches it

We build MCP servers and tool connections as governed integrations: scoped credentials, approved tools only, and audit logging, so each agent gets the access it needs and no more.

Get Started with Agentic AI →

Related terms.

Architecture

Model Context Protocol (MCP)

MCP and A2A

In the IBM stack

How Incede.ai approaches it

See also on Incede.ai

Related terms.

Agent2Agent (A2A) protocol

AI agent

IBM Bob

Agent control plane

Ready to put Model Context Protocol (MCP) to work?