A2A Protocol: How Agents Discover and Collaborate with Other Agents

~15 min read

A2A lets AI agents connect to and collaborate with OTHER agents — distinct from MCP, which connects agents to tools. Agents publish a JSON "Agent Card" describing their capabilities so others can discover and communicate with them.

A2A Protocol: How Agents Discover and Collaborate with Other Agents is a Pro topic

Sign in, then upgrade to Pro or Power to unlock this topic and the full AI Engineering curriculum.

Key points

  • A2A lets agents connect with and collaborate with OTHER agents; MCP lets agents connect with tools/APIs — different, complementary roles
  • Agents collaborate by exchanging context, task updates, instructions, and data — WITHOUT sharing internal memory, thoughts, or tools
  • Remote agents publish a JSON 'Agent Card' detailing capabilities and authentication, enabling discovery
  • AI applications can model A2A agents as MCP resources, represented by their Agent Card
  • Key strength: agents built on entirely different frameworks (LlamaIndex, CrewAI, etc.) can still collaborate through this standard