Agent-to-Agent Protocol (A2A)
Definition
A communication protocol, developed by Google, that enables autonomous agents to discover, negotiate with, and delegate tasks to other agents — without human mediation. Unlike MCP (which connects models to tools), A2A connects agents to agents. It supports capability discovery via Agent Cards (JSON metadata at /.well-known/agent.json), task lifecycle management (submitted → working → completed/failed), streaming updates via SSE, and multi-turn conversations between agents. A2A is designed for asynchronous, long-running tasks where agents collaborate as peers.
Builder Context
A2A matters when your agent needs to delegate work to another agent that may take minutes or hours — not milliseconds. Think: 'research this topic and get back to me' or 'generate a report from these 5 data sources.' MCP handles the tool calls; A2A handles the agent collaboration. In practice, most builders start with MCP and add A2A when they need multi-agent orchestration. The Agent Card discovery mechanism is useful even without full A2A adoption — it gives your agent a machine-readable capability manifest. A2A is still early (2025 launch) but backed by 50+ partners including Salesforce, SAP, and LangChain.