AI agent glossary
Agent2Agent (A2A) protocol
An open protocol that lets AI agents built by different vendors discover each other, exchange messages, and delegate tasks across systems.
MCP connects an agent to tools and data. A2A, short for Agent2Agent, addresses a different problem: letting one agent work with another agent that someone else built and runs. Google introduced it in 2025 and later contributed it to the Linux Foundation.
In A2A, an agent publishes an Agent Card, a JSON document describing what it can do, how to reach it, and what authentication it needs. A client agent can then send it tasks and receive progress updates and results, including for long-running work.
A2A and MCP are complementary rather than competing. A practical rule: if the other side is a tool or data source you call, MCP fits; if it is an autonomous agent that plans its own steps, A2A is designed for that case. Most builders will need MCP long before they need A2A.
Learn more in the course
Related terms
Go from definitions to doing
Learn to use and build AI agents in 65 practical lessons.