What is A2A (Agent2Agent Protocol)?
Updated May 2026
A2A is the protocol that lets two AI agents talk to each other — for example, one calendar agent scheduling with another calendar agent.
Google introduced the Agent2Agent Protocol in 2025 as the spec for direct agent-to-agent communication. While MCP defines how an AI client talks to a tool server, A2A defines how two agents (each potentially with their own tools, voice, and identity) negotiate a task together.
A typical A2A scenario: a user's personal AI agent reaches out to a creator's withlove agent to book a meeting. The two agents negotiate the time, share the relevant context, and confirm the booking — all without the human typing.
Every withlove agent has A2A support on the roadmap. The platform already ships did:web identifiers and Agent Cards that A2A clients can discover.
Frequently asked questions
How is A2A different from MCP?
MCP is client-to-server: an AI talks to a tool. A2A is agent-to-agent: two AI agents negotiate a task together. They complement each other; most agents will expose both surfaces.
Related terms
Your own AI assistant in five minutes
Skip the jargon. withlove gives you a working AI agent at a shareable URL — voice, calendar, prices, leads — without any of the setup.
Try withlove free